# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/customInferenceImages/{imageId}/featureImpact/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
imageId
path
string
true
ID of the image of the custom inference model to retrieve feature impact from.
Cannot retrieve feature impact scores: (1) if custom model is not an inference model, (2) if training data is not assigned, (3) if feature impact job is in progress for custom model.
None
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/customInferenceImages/{imageId}/featureImpact/¶
Add a request to calculate feature impact for a custom inference model image to the queue.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/customInferenceImages/{imageId}/featureImpact/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
If job cannot be submitted because of invalid input or model state: (1) if image id does not correspond to a custom inference model, (2) if training data is not yet assigned or assignment is in progress, (3) if the rowCount exceeds the minimum or maximum value for this model's training data.
# You can also use wget
curl-XDELETEhttps://app.datarobot.com/api/v2/customModels/{customModelId}/versions/{customModelVersionId}/dependencyBuild/\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/customModels/{customModelId}/versions/{customModelVersionId}/dependencyBuild/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/customModels/{customModelId}/versions/{customModelVersionId}/dependencyBuild/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/customModels/{customModelId}/versions/{customModelVersionId}/dependencyBuildLog/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/customTrainingBlueprints/?offset=0&limit=1000\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
true
This many results will be skipped.
limit
query
integer
true
At most this many results are returned.
customModelId
query
string
false
List blueprints for a specific model. Default: all.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/customTrainingBlueprints/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{CustomTrainingBlueprintCreate}'
Input parameters are invalid: either the custom model is for inference or no environment version ID was specified and the given environment has no versions.
None
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/executionEnvironments/{environmentId}/accessControl/¶
Get a list of users who have access to this execution environment and their roles.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/executionEnvironments/{environmentId}/accessControl/?offset=0&limit=0\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
true
This many results will be skipped
limit
query
integer
true
At most this many results are returned
username
query
string
false
Optional, only return the access control information for a user with this username.
userId
query
string
false
Optional, only return the access control information for a user with this user ID.
# You can also use wget
curl-XPATCHhttps://app.datarobot.com/api/v2/executionEnvironments/{environmentId}/accessControl/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/modelPackages/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
This many results will be skipped.
limit
query
integer
false
At most this many results are returned.
modelId
query
string
false
If specified, limit results to model packages for the model with the specified ID.
similarTo
query
string
false
Return model packages similar to a given model package ID. If used, will only return model packages that match target.name, target.type, target.classNames (for classification models), modelKind.isTimeSeries, and modelKind.isMultiseries of the specified model package.
forChallenger
query
boolean
false
Can be used with similarTo to request similar model packages with the intent to use them as challenger models; for external model packages, instead of returning similar external model packages, similar DataRobot and Custom model packages will be retrieved.
search
query
string
false
Provide a term to search for in package name, model name, or description
predictionThreshold
query
number
false
Prediction threshold used for binary classification models
imported
query
boolean
false
If specified, filter for either imported (true) or non-imported (false) model packages
predictionEnvironmentId
query
string
false
Can be used to filter packages by what is supported by the prediction environment
modelKind
query
any
false
Return models from the registry that match a specific format.
buildStatus
query
string
false
If specified, filter model packages by the build status.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/modelPackages/fromLeaderboard/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{ModelPackageCreateFromLeaderboard}'
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/modelPackages/fromLearningModel/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{ModelPackageCreateFromLearningModel}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/modelPackages/{modelPackageId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Either the model package does not exist or the user does not have permission to view the model package.
None
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/modelPackages/{modelPackageId}/archive/¶
(Deprecated in v2.32) Permanently archive a model package. It will no longer be able to be used in new deployments or replacement. It will not be accessible in the model package list api. It will only be accessible at the model package retrieve route for this model package.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/modelPackages/{modelPackageId}/capabilities/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/modelPackages/{modelPackageId}/features/?offset=0&limit=50\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
true
The number of features to skip, defaults to 0.
limit
query
integer
true
The number of features to return, defaults to 0.
includeNonPredictionFeatures
query
string
false
When True will return all raw features in the universe dataset associated with the deployment, and when False will return only those raw features used to make predictions on the deployment.
forSegmentedAnalysis
query
string
false
When True, features returned will be filtered to those usable for segmented analysis.
search
query
string
false
Case insensitive search against names of the deployment's features.
orderBy
query
string
false
The sort order to apply to the list of features. Prefix the attribute name with a dash to sort in descending order, e.g., "-name".
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/modelPackages/{modelPackageId}/modelLogs/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/modelPackages/{modelPackageId}/sharedRoles/¶
(Deprecated in v2.32) instead.Get a list of users, groups and organizations who have access to this model package and their roles on the model package.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/modelPackages/{modelPackageId}/sharedRoles/?offset=0&limit=10\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
id
query
string
false
Only return roles for a user, group or organization with this identifier.
offset
query
integer
true
This many results will be skipped
limit
query
integer
true
At most this many results are returned
name
query
string
false
Only return roles for a user, group or organization with this name.
shareRecipientType
query
string
false
List access controls for recipients with this type.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/biasMitigatedModels/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
This many results will be skipped.
limit
query
integer
false
At most this many results are returned.
parentModelId
query
string
false
Retrieve a list of mitigated models for the parent model if specified. Otherwise retrieve a list of all mitigated models for the project.
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/projects/{projectId}/biasMitigatedModels/¶
Add a request to the queue to train a model with bias mitigation applied.
If the job has been previously submitted, the request will return the jobId of the previously submitted job. Use this jobId to check status of the previously submitted job.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/biasMitigatedModels/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/biasMitigationFeatureInfo/?featureName=string\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/biasMitigationFeatureInfo/{featureName}/\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/blenderModels/?offset=0&limit=0\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
true
This many results will be skipped.
limit
query
integer
true
At most this many results are returned. If 0, all results.
Create a blender from other models using a specified blender method. Note: Time Series projects only allow the following blender methods: "AVG", "MED", "FORECAST_DISTANCE_ENET", and "FORECAST_DISTANCE_AVG".
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/blenderModels/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/blenderModels/blendCheck/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/blenderModels/{modelId}/¶
Retrieve a blender. Blenders are a special type of models, so the response includes all attributes that would be in a response to GET /api/v2/projects/{projectId}/models/{modelId}/ as well as some additional ones.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/blenderModels/{modelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Retrieve all existing combined models for this project.
.. note::
To retrieve information on the segments for a combined model, retrieve the combined model using [GET /api/v2/projects/{projectId}/combinedModels/{combinedModelId}/][get-apiv2projectsprojectidcombinedmodelscombinedmodelid]
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/combinedModels/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
Number of results to skip.
limit
query
integer
false
At most this many results are returned. The default may change without notice.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/combinedModels/{combinedModelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/combinedModels/{combinedModelId}/segments/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
Number of results to skip.
limit
query
integer
false
At most this many results are returned. The default may change without notice.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/combinedModels/{combinedModelId}/segments/download/\-H"Accept: text/csv"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/datetimeModels/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
Number of results to skip.
limit
query
integer
false
At most this many results are returned. The default may change without notice.
bulkOperationId
query
string
false
the ID of the bulk model operation. If specified, only models submitted in scope of this operation will be shown.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/datetimeModels/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/projects/{projectId}/datetimeModels/fromModel/¶
Retrain an existing datetime model using a new training period for the model training set (with optional time window sampling) or different feature list.
All durations and datetimes should be specified in accordance with the :ref:timestamp and duration formatting rules<time_format>.
Note that only one of trainingDuration or trainingRowCount or trainingStartDate and trainingEndDate should be specified. If trainingStartDate and trainingEndDate are specified, the source model must be frozen.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/datetimeModels/fromModel/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/datetimeModels/{modelId}/¶
Look up a particular datetime model
All durations and datetimes are specified in accordance with :ref:timestamp and duration formatting rules <time_format>.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/datetimeModels/{modelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/datetimeModels/{modelId}/backtests/\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/deploymentReadyModels/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/eureqaDistributionPlot/{solutionId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/eureqaModelDetail/{solutionId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/eureqaModels/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/eureqaModels/{modelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/projects/{projectId}/frozenDatetimeModels/¶
Train a frozen datetime model. If no training data is specified, the frozen datetime model will be trained on the most recent data using an amount of data that is equivalent to the original model. However, if the equivalent duration does not provide enough rows for training, then the duration will be extended until the minimum is met. Note that this will require the holdout of the project to be unlocked.
All durations and datetimes should be specified in accordance with the :ref:timestamp and duration formatting rules<time_format>.
Note that only one of trainingDuration, trainingRowCount, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/frozenDatetimeModels/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/frozenModels/?offset=0&limit=0\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
true
This many results will be skipped.
limit
query
integer
true
At most this many results are returned. If 0, all results.
withMetric
query
string
false
If specified, the returned models will only have scores for this metric. If not, all metrics will be included.
Train a new frozen model with parameters from an existing model. Frozen models use tuning parameters from another model on the leaderboard, allowing them to be retrained on a larger amount of the training data more efficiently.
To specify the amount of data to use to train the model, use either samplePct to express a percentage of the rows of the dataset to use or trainingRowCount to express the number of rows to use.
If neither samplePct or trainingRowCount is specified, the model will be trained on the maximum available training data that can be used to train an in-memory model.
For projects using smart sampling, samplePct and trainingRowCount will be interpreted as a percent or number of rows of the minority class.
When configuring retraining sample sizes for models in projects with large row counts, DataRobot recommends requesting sample sizes using integer row counts instead of percentages. This is because percentages map to many actual possible row counts and only one of which is the actual sample size for up to validation. For example, if a project has 199,408 rows and you request a 64% sample size, any number of rows between 126,625 rows and 128,618 rows maps to 64% of the data. Using actual integer row counts (or project.max_training_rows) avoids ambiguity around how many rows of data you want the model to use.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/frozenModels/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/frozenModels/{modelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/incrementalLearningModels/fromModel/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/modelJobs/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
status
query
string
false
If provided, only jobs with the same status will be included in the results; otherwise, queued and inprogress jobs (but not errored jobs) will be returned.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/modelJobs/{jobId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/modelRecords/?offset=0&limit=100\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
withMetric
query
string
false
If specified, the returned models will only have scores for this metric. If not, all metrics will be included.
showInSampleScores
query
boolean
false
If specified, will return metric scores for models trained into validation/holdout for projects that do not have stacked predictions.
characteristics
query
array[string]
false
A characteristics to filter models with. Only those models which have all specified characteristics are returned.
searchTerm
query
string
false
Filter models by the string expression in the description, case insensitive.
labels
query
array[string]
false
Filter models by labels. Only those models which have all specified labels are returned.
blueprints
query
string
false
Filter models by blueprint ids.
families
query
array[string]
false
Filter models by families.
featurelists
query
string
false
Filter models by featurelist names.
trainingFilters
query
string
false
Filter models by training length or type. Could be training duration in string representation, 'Start/end date', 'Project settings' constants or number of rows. Training duration for datetime partitioned projects may have up to three parts: --.Example of the training duration: P6Y0M0D-78-Random, returns models trained on6 years, with sampling rate 78%, randomly taken from the training window. Example of the row count with sampling method: 150-Random
numberOfClusters
query
string
false
Filter models by number of clusters. Applicable only in unsupervised clustering projects.
sortByMetric
query
string
false
Metric to order models by. If omitted, the project metric is used.
sortByPartition
query
string
false
Partition to use when ordering models by metric. If omitted, the validation partition is used.
Use [GET /api/v2/projects/{projectId}/modelRecords/][get-apiv2projectsprojectidmodelrecords] instead.
Fewer attributes are returned in the response of model records route.
Removed attributes:
monotonic_increasing_featurelist_id -- Retrievable from the blueprint level
monotonic_decreasing_featurelist_id -- Retrievable from the blueprint level.
supports_composable_ml -- Retrievable from the blueprint level.
supports_monotonic_constraints -- Retrievable from the blueprint level.
has_empty_clusters -- Retrievable from the individual model level.
is_n_clusters_dynamically_determined -- Retrievable from the individual model level.
prediction_threshold -- Retrievable from the individual model level.
prediction_threshold_read_only - Retrievable from the individual model level.
Changed attributes:
n_clusters becomes number_of_clusters and is returned for unsupervised clustering models.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
withMetric
query
string
false
If specified, the returned models will only have scores for this metric. If not, all metrics will be included.
showInSampleScores
query
boolean
false
If specified, will return metric scores for models trained into validation/holdout for projects that do not have stacked predictions.
name
query
string
false
If specified, filters for models with a model type matching name.
samplePct
query
number
false
If specified, filters for models with a matching sample percentage.
isStarred
query
string
false
If specified, filters for models marked as starred.
orderBy
query
string
false
A comma-separated list of metrics to sort by. If metric is prefixed with a '-', models are sorted by this metric in descending order, otherwise are sorted in ascending order. Valid sorting metrics are metric and samplePct. Use of metric sorts models by metric value selected for this project using the validation score. Use of the prefix accounts for the direction of the metric, so -metric will sort in order of decreasing 'goodness', which may be opposite to the natural numerical order. If not specified, -metric will be used.
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
» featurelistId
string¦null
true
The ID of the feature list used by the model.
» featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
» hasFinetuners
boolean
false
Whether a model has fine tuners.
» id
string
true
The ID of the model.
» isAugmented
boolean
false
Whether a model was trained using augmentation.
» isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
» isNClustersDynamicallyDetermined
boolean
false
Whether number of clusters is dynamically determined. Only valid in unsupervised clustering projects.
» isStarred
boolean
true
Indicates whether the model has been starred.
» isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
» isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
Reason for lifecycle stage. None if model is active.
»» stage
string
true
Model lifecycle stage.
» linkFunction
string¦null
true
The link function the final modeler uses in the blueprint. If no link function exists, returns null.
» metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
» modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
» modelFamily
string
true
the family model belongs to, e.g. SVM, GBM, etc.
» modelFamilyFullName
string
true
The full name of the family that the model belongs to. For e.g., Support Vector Machine, Gradient Boosting Machine, etc.
» modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
» modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
» monotonicDecreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
» monotonicIncreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
» nClusters
integer¦null
false
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
» parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
» predictionThreshold
number
true
maximum: 1 minimum: 0
threshold used for binary classification in predictions.
» predictionThresholdReadOnly
boolean
true
indicates whether modification of a predictions threshold is forbidden. Since v2.22 threshold modification is allowed.
» processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
» projectId
string
true
The ID of the project to which the model belongs.
» samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
» samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
» supportsComposableMl
boolean
true
indicates whether this model is supported in Composable ML.
» supportsMonotonicConstraints
boolean
true
whether this model supports enforcing monotonic constraints
» timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
» trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
» trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
» trainingRowCount
integer¦null
true
minimum: 0 (exclusive)
The number of rows used to train the model.
» trainingStartDate
string(date-time)¦null
true
the start date of the dates in the partition column for the data used to train the model
Train a new model. To specify the amount of data to use to train the model, use either samplePct to express a percentage of the rows of the dataset to use or trainingRowCount to express the number of rows to use. If neither samplePct or trainingRowCount is specified, the model will be trained on the maximum available training data that can be used to train an in-memory model. For projects using smart sampling, samplePct and trainingRowCount will be interpreted as a percent or number of rows of the minority class.
When configuring retraining sample sizes for models in projects with large row counts, DataRobot recommends requesting sample sizes using integer row counts instead of percentages. This is because percentages map to many actual possible row counts and only one of which is the actual sample size for up to validation. For example, if a project has 199,408 rows and you request a 64% sample size, any number of rows between 126,625 rows and 128,618 rows maps to 64% of the data. Using actual integer row counts (or project.max_training_rows) avoids ambiguity around how many rows of data you want the model to use.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/models/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/projects/{projectId}/models/fromModel/¶
Retrain an existing model using a new sample size and/or feature list.When configuring retraining sample sizes for models in projects with large row counts, DataRobot recommends requesting sample sizes using integer row counts instead of percentages. This is because percentages map to many actual possible row counts and only one of which is the actual sample size for up to validation. For example, if a project has 199,408 rows and you request a 64% sample size, any number of rows between 126,625 rows and 128,618 rows maps to 64% of the data. Using actual integer row counts (or project.max_training_rows) avoids ambiguity around how many rows of data you want the model to use.
Note that only one of samplePct or trainingRowCount should be specified.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/models/fromModel/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPATCHhttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/projects/{projectId}/models/{modelId}/advancedTuning/¶
Submit a job to make a new version of the model with different advanced tuning parameters. Note: This route currently supports all models other than: OSS, blenders, prime, scaleout, baseline and user-created. Currently, only single-stage models (most simple models) are supported. Blueprints that run multiple steps, for example one step to predict zero vs nonzero and a second step to determine the value of nonzero predictions, are not supported. (:ref:Advanced Tuning documentation <grid_search>). Parameters may be omitted from this endpoint. If a parameter is omitted, its currentValue will be used. To see the possible parameter IDs and constraints on possible values, see GET /api/v2/projects/{projectId}/models/{modelId}/advancedTuning/parameters/.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/advancedTuning/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
A url at which the job processing the model can be retrieved.
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/models/{modelId}/advancedTuning/parameters/¶
Retrieve information about all advanced tuning parameters available for the specified model. Note: This route currently supports all models other than: OSS, blenders, prime, scaleout, baseline and user-created
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/advancedTuning/parameters/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/clusterNames/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Could not find unsupervised clustering model. Possible reasons include: 1. Provided model id points to a model that does not exist in specified project. 2. Provided model has incompatible type. Method requires model to be unsupervised clustering model.
None
To perform this operation, you must be authenticated by means of one of the following methods:
# You can also use wget
curl-XPATCHhttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/clusterNames/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
Could not find unsupervised clustering model. Possible reasons include: 1. Provided model id points to a model which does not exists in specified project. 2. Provided model has incompatible type. Method requires model to be unsupervised clustering model.
The request cannot be processed. Possible reasons include: 1. Mapping contains invalid current cluster name and referenced cluster was not found. 2. Mapping is invalid as after update, clusters will not be uniquely identifiable by name.
None
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/projects/{projectId}/models/{modelId}/crossValidation/¶
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/crossValidation/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
Contains a url at which the job processing the model can be retrieved
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/models/{modelId}/crossValidationScores/¶
Get Cross Validation scores for each partition in a model.
.. note:: Individual partition scores are only available for newer models; older models that
have cross validation score calculated will need to be retrained.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/crossValidationScores/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
metric
query
string
false
Set to the name of a metric to only return results for that metric.
partition
query
number
false
Set to a value such as 1.0, 2.0 to only return results for the specified partition.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/features/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/models/{modelId}/missingReport/¶
Retrieve a summary of how the model's subtasks handle missing values
Only models built after the missing value report feature was added will have reports,
and only models with at least one imputation or encoding task, e.g. ordinal encoding,
missing value imputation. Blenders and scaleout models do not support Missing Value reports.
The report will describe how each feature's missing values were treated, and report how many
missing values were present in the training data. Features which were not processed by a
given blueprint task will not mention it: for instance, a categorical feature with many
unique values may not be considered eligible for processing by a One-Hot Encoding
Report is collected for those features which are considered eligible by given
blueprint task. For instance, categorical feature with a lot of unique values may not be
considered as eligible in One-Hot Encoding Task.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/missingReport/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/numIterationsTrained/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Cannot retrieve early stopping information for this model.
None
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/models/{modelId}/parameters/¶
Retrieve model parameters. These are the parameters that appear in the webapp on the Coefficients tab. Note that they are only available for some models.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/parameters/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/models/{modelId}/predictionIntervals/¶
Retrieve prediction intervals (in descending order) that are already calculated for this model.
Note that the project this model belongs to must be a time series project.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/predictionIntervals/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
This many results will be skipped.
limit
query
integer
false
At most this many results are returned.
projectId
path
string
true
The project to retrieve prediction intervals for. Must be a time series project.
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
POST /api/v2/projects/{projectId}/models/{modelId}/predictionIntervals/¶
Submit a job to calculate prediction intervals for the specified percentiles for this model.
Note that the project this model belongs to must be a time series project.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/predictionIntervals/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/primeInfo/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/models/{modelId}/primeRulesets/¶
List all the rulesets approximating a model
When rulesets are created for the parent model, all of the rulesets are created at once, but not all rulesets have corresponding Prime models (until they are directly requested).
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/primeRulesets/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/primeRulesets/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
a url that can be polled to check the status of the job
To perform this operation, you must be authenticated by means of one of the following methods:
BearerAuth
GET /api/v2/projects/{projectId}/models/{modelId}/scoringCode/¶
Retrieve Scoring Code for making new predictions from an existing model offline.
You need the "Scoring Code" feature enabled to use this route.
By default, returns a compiled executable JAR that can be executed locally to calculate model predictions, or it can be used as a library for a Java application. Execute it with the '--help` parameters to learn how to use it as a command-line utility.
See model API documentation (https://javadoc.io/doc/com.datarobot/datarobot-prediction/latest/index.html) to be able to use it inside an existing Java application.
With the sourceCode query parameter set to 'true', returns a source code archive that can be used to review internal calculations of the model. This JAR is NOT executable.
See "https://docs.datarobot.com/en/docs/predictions/port-pred/scoring-code/index.html" in DataRobot application for more information.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/scoringCode/\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
sourceCode
query
string
false
If set to "true", the downloaded JAR file will contain only the source code and will not be executable.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/models/{modelId}/supportedCapabilities/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/primeFiles/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
this many results will be skipped
limit
query
integer
false
at most this many results are returned. To specify no limit, use 0. The default may change and a maximum limit may be imposed without notice.
parentModelId
query
string
false
if specified only Prime files approximating the specified parent model will be returned; otherwise all applicable Prime files will be returned.
modelId
query
string
false
if specified only Prime files with code used in the specified prime model will be returned; otherwise all applicable Prime files will be returned
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/primeFiles/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/primeFiles/{primeFileId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/primeFiles/{primeFileId}/download/\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/primeModels/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
Number of results to skip.
limit
query
integer
false
At most this many results are returned. The default may change without notice.
DataRobot Prime is not available for multiclass projects.
Once rulesets approximating a parent model have been created, using POST /api/v2/projects/(projectId)/models/(modelId)/primeRulesets/, this route will allow creation of a Prime model using one of those rulesets.
Available rulesets can be retrieved via GET /api/v2/projects/(projectId)/models/(modelId)/primeRulesets/. Deprecated in v2.35.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/primeModels/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/primeModels/{modelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Prime models are an extension of models, so the response includes all attributes that would be in a response to GET /api/v2/projects/(projectId)/models/(modelId)/ as well as some additional ones.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTableModels/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
withMetric
query
string
false
If specified, the returned models will only have scores for this metric. If not, all metrics will be included.
showInSampleScores
query
boolean
false
If specified, will return metric scores for models trained into validation/holdout for projects that do not have stacked predictions.
name
query
string
false
If specified, filters for models with a model type matching name.
samplePct
query
number
false
If specified, filters for models with a matching sample percentage.
isStarred
query
string
false
If specified, filters for models marked as starred.
orderBy
query
string
false
A comma-separated list of metrics to sort by. If metric is prefixed with a '-', models are sorted by this metric in descending order, otherwise are sorted in ascending order. Valid sorting metrics are metric and samplePct. Use of metric sorts models by metric value selected for this project using the validation score. Use of the prefix accounts for the direction of the metric, so -metric will sort in order of decreasing 'goodness', which may be opposite to the natural numerical order. If not specified, -metric will be used.
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
» featurelistId
string¦null
true
The ID of the feature list used by the model.
» featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
» hasFinetuners
boolean
false
Whether a model has fine tuners.
» id
string
true
The ID of the model.
» isAugmented
boolean
false
Whether a model was trained using augmentation.
» isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
» isNClustersDynamicallyDetermined
boolean
false
Whether number of clusters is dynamically determined. Only valid in unsupervised clustering projects.
» isStarred
boolean
true
Indicates whether the model has been starred.
» isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
» isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
Reason for lifecycle stage. None if model is active.
»» stage
string
true
Model lifecycle stage.
» linkFunction
string¦null
true
The link function the final modeler uses in the blueprint. If no link function exists, returns null.
» metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
» modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
» modelFamily
string
true
the family model belongs to, e.g. SVM, GBM, etc.
» modelFamilyFullName
string
true
The full name of the family that the model belongs to. For e.g., Support Vector Machine, Gradient Boosting Machine, etc.
» modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
» modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
» monotonicDecreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
» monotonicIncreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
» nClusters
integer¦null
false
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
» parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
» predictionThreshold
number
true
maximum: 1 minimum: 0
threshold used for binary classification in predictions.
» predictionThresholdReadOnly
boolean
true
indicates whether modification of a predictions threshold is forbidden. Since v2.22 threshold modification is allowed.
» processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
» projectId
string
true
The ID of the project to which the model belongs.
» ratingTableId
string
true
The rating table ID
» samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
» samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
» supportsComposableMl
boolean
true
indicates whether this model is supported in Composable ML.
» supportsMonotonicConstraints
boolean
true
whether this model supports enforcing monotonic constraints
» timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
» trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
» trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
» trainingRowCount
integer¦null
true
minimum: 0 (exclusive)
The number of rows used to train the model.
» trainingStartDate
string(date-time)¦null
true
the start date of the dates in the partition column for the data used to train the model
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTableModels/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTableModels/{modelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
List RatingTable objects for a project.
These contain metadata about the rating table and the location at which the corresponding rating
table file can be retrieved.
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTables/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
parentModelId
query
string
false
optional. If specified, only rating tables with this parentModelId will be returned
modelId
query
string
false
optional. If specified, only rating tables with this modelId will be returned
offset
query
integer
false
optional (default: 0), this many results will be skipped
limit
query
integer
false
optional (default: no limit), at most this many results are returned. To specify no limit, use 0. The default may change and a maximum limit may be imposed without notice.
Create a new rating table from a rating table file
This will create a new rating table, regardless of whether the validation succeeds.
The rating table object will have a validationError which will be left blank in the case of successful validation.
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTables/\-H"Content-Type: multipart/form-data"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTables/{ratingTableId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
projectId
path
string
true
the project that owns this data
ratingTableId
path
string
true
the rating table ID to retrieve the source file from
# You can also use wget
curl-XPATCHhttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTables/{ratingTableId}/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ratingTables/{ratingTableId}/file/\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
projectId
path
string
true
the project that owns this data
ratingTableId
path
string
true
the rating table ID to retrieve the source file from
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/recommendedModels/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/recommendedModels/recommendedModel/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ruleFitFiles/?offset=0&limit=100\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
modelId
query
string
false
if specified only RuleFit code files with code used in the specified RuleFit model will be returned; otherwise all applicable RuleFit files will be returned
# You can also use wget
curl-XPOSThttps://app.datarobot.com/api/v2/projects/{projectId}/ruleFitFiles/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ruleFitFiles/{ruleFitFileId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/projects/{projectId}/ruleFitFiles/{ruleFitFileId}/download/\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPUThttps://app.datarobot.com/api/v2/projects/{projectId}/segmentChampion/\-H"Content-Type: application/json"\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/registeredModels/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
This many results will be skipped.
limit
query
integer
false
At most this many results are returned.
search
query
string
false
A term to search for in registered model name
createdAtStartTs
query
string(date-time)
false
Registered models created on or after this timestamp
createdAtEndTs
query
string(date-time)
false
Registered models created before this timestamp. Defaults to the current time
modifiedAtStartTs
query
string(date-time)
false
Registered models modified on or after this timestamp
modifiedAtEndTs
query
string(date-time)
false
Registered models modified before this timestamp. Defaults to the current time
targetName
query
string
false
Name of the target to filter by
targetType
query
string
false
Type of the target to filter by
createdBy
query
string
false
Email of the user that created registered model to filter by
sortKey
query
string
false
Key to order result by
sortDirection
query
string
false
Sort direction
compatibleWithLeaderboardModelId
query
string
false
If specified, limit results to registered models containing versions (model packages) for the leaderboard model with the specified ID.
compatibleWithModelPackageId
query
string
false
Return registered models that have versions (model packages) compatible with given model package ID. If used, will only return registered models which have versions that match target.name, target.type, target.classNames (for classification models), modelKind.isTimeSeries, and modelKind.isMultiseries of the specified model package.
forChallenger
query
boolean
false
Can be used with compatibleWithModelPackageId to request similar registered models that contain versions (model packages)that can be used as challenger models; for external model packages, instead of returning similar external model packages, similar DataRobot and Custom model packages will be retrieved.
predictionThreshold
query
number
false
If specified, return any registered models containing one or more versions matching the prediction threshold used for binary classification models
imported
query
boolean
false
If specified, return any registered models that contain either imported (true) or non-imported (false) versions (model packages)
predictionEnvironmentId
query
string
false
Can be used to filter registered models by what is supported by the prediction environment
modelKind
query
any
false
Return models that contain versions matching a specific format
buildStatus
query
string
false
If specified, only return models that have versions with specified build status
stage
query
string
false
If specified, only returns models that have versions in the specified stage.
isGlobal
query
boolean
false
Return only global (accessible to all users in the organization) registered models or local(accessible only to the owner and the users with whom it has been explicitly shared)
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XPATCHhttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/deployments/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/sharedRoles/?offset=0&limit=10\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
id
query
string
false
Only return roles for a user, group or organization with this identifier.
offset
query
integer
true
This many results will be skipped
limit
query
integer
true
At most this many results are returned
name
query
string
false
Only return roles for a user, group or organization with this name.
shareRecipientType
query
string
false
List access controls for recipients with this type.
# You can also use wget
curl-XPATCHhttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/sharedRoles/\-H"Content-Type: application/json"\-H"Authorization: Bearer {access-token}"\-d'{undefined}'
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/versions/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Parameters
Name
In
Type
Required
Description
offset
query
integer
false
This many results will be skipped.
limit
query
integer
false
At most this many results are returned.
sortKey
query
string
false
Key to order result by
sortDirection
query
string
false
Sort direction
targetName
query
string
false
Name of the target to filter by
targetType
query
string
false
Type of the target to filter by
search
query
string
false
A term to search for in version name, model name, or description
compatibleWithLeaderboardModelId
query
string
false
If specified, limit results to versions (model packages) of the leaderboard model with the specified ID.
compatibleWithModelPackageId
query
string
false
Return versions compatible with given model package ID. If used, will only return versions that match target.name, target.type, target.classNames (for classification models), modelKind.isTimeSeries, and modelKind.isMultiseries of the specified model package
forChallenger
query
boolean
false
Can be used with compatibleWithModelPackageId to request similar versions that can be used as challenger models; for external model packages, instead of returning similar external model packages, similar DataRobot and Custom model packages will be retrieved
predictionThreshold
query
number
false
Return versions with the specified prediction threshold used for binary classification models
imported
query
boolean
false
If specified, return either imported (true) or non-imported (false) versions (model packages)
predictionEnvironmentId
query
string
false
Can be used to filter versions (model packages) by what is supported by the prediction environment
modelKind
query
any
false
Return versions that match a specific format.
buildStatus
query
string
false
If specified, filter versions by the build status.
stage
query
string
false
If specified, filter versions by the stage.
useCaseId
query
string
false
If specified, filter versions by use-case id.
createdBy
query
string
false
Email of the user that created registered model version to filter by
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/versions/{versionId}/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
# You can also use wget
curl-XGEThttps://app.datarobot.com/api/v2/registeredModels/{registeredModelId}/versions/{versionId}/deployments/\-H"Accept: application/json"\-H"Authorization: Bearer {access-token}"
Indicates that the value can contain free-form ASCII text. If present, is an empty object. Note that ascii fields must be valid ASCII-encoded strings. Additionally, they may not contain semicolons or newlines.
Specifies whether the mitigation feature will be used as a predictor variable (i.e., treated like other categorical features in the input to train the modeler), in addition to being used for bias mitigation. If false, the mitigation feature will be used only for bias mitigation, and not for training the modeler task.
modelId
string
true
Mitigated model ID
parentModelId
string¦null
true
Parent model ID
protectedFeature
string
true
Protected feature that will be used in a bias mitigation task to mitigate bias
The name of the protected feature used to mitigate bias on models.
biasMitigationParentLid
string
true
The ID of the model to modify with a bias-mitigation task.
biasMitigationTechnique
string
true
Method applied to perform bias mitigation.
includeBiasMitigationFeatureAsPredictorVariable
boolean
true
Specifies whether the mitigation feature will be used as a predictor variable (i.e., treated like other categorical features in the input to train the modeler), in addition to being used for bias mitigation. If false, the mitigation feature will be used only for bias mitigation, and not for training the modeler task.
The blender method, one of "PLS", "GLM", "AVG", "ENET", "MED", "MAE", "MAEL1", "TF", "RF", "LGBM", "FORECAST_DISTANCE_ENET" (new in v2.18), "FORECAST_DISTANCE_AVG" (new in v2.18), "MIN", "MAX".
Method used to blend results of underlying models.
blueprintId
string
true
The blueprint used to construct the model.
dataSelectionMethod
string
false
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
featurelistId
string¦null
true
The ID of the feature list used by the model.
featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
hasFinetuners
boolean
false
Whether a model has fine tuners.
id
string
true
The ID of the model.
isAugmented
boolean
false
Whether a model was trained using augmentation.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isNClustersDynamicallyDetermined
boolean
false
Whether number of clusters is dynamically determined. Only valid in unsupervised clustering projects.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
The link function the final modeler uses in the blueprint. If no link function exists, returns null.
metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
the family model belongs to, e.g. SVM, GBM, etc.
modelFamilyFullName
string
true
The full name of the family that the model belongs to. For e.g., Support Vector Machine, Gradient Boosting Machine, etc.
modelIds
[string]
true
List of models used in blender.
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
monotonicDecreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
monotonicIncreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
nClusters
integer¦null
false
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
predictionThreshold
number
true
maximum: 1 minimum: 0
threshold used for binary classification in predictions.
predictionThresholdReadOnly
boolean
true
indicates whether modification of a predictions threshold is forbidden. Since v2.22 threshold modification is allowed.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
supportsComposableMl
boolean
true
indicates whether this model is supported in Composable ML.
supportsMonotonicConstraints
boolean
true
whether this model supports enforcing monotonic constraints
timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
trainingRowCount
integer¦null
true
minimum: 0 (exclusive)
The number of rows used to train the model.
trainingStartDate
string(date-time)¦null
true
the start date of the dates in the partition column for the data used to train the model
The end of the numeric range for the current bin. Note that binEnd - binStart should be a constant, modulo floating-point rounding error, for all bins in a single plot.
binStart
number
true
The start of the numeric range for the current bin. Must be equal to the binEnd of the previous bin.
negatives
integer
true
The number of records in the dataset where the model's predicted value falls into this bin and the target is negative.
positives
integer
true
The number of records in the dataset where the model's predicted value falls into this bin and the target is positive.
A list of the mappings from a cluster's current name to its new name. After update, value passed as a new name will become cluster's current name. All cluster names should be unique and should identify one and only one cluster.
Indicates whether the segment champion model is frozen, i.e. uses tuning parameters from a parent model
modelAssignedBy
string¦null
true
Who assigned model as segment champion. Default is DataRobot.
modelAwardTime
string(date-time)¦null
true
Time when model was awarded as segment champion.
modelCount
integer¦null
true
Count of trained models in project.
modelIcon
[integer]¦null
true
The number for the icon representing the given champion model.
modelId
string¦null
true
ID of segment champion model.
modelMetrics
object¦null
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelType
string¦null
true
The description of the model type of the given champion model.
projectId
string¦null
true
The ID of the project.
projectPaused
boolean¦null
false
Is project paused right now.
projectStage
string¦null
true
The current stage of the project, where modeling indicates that the target has been successfully set and modeling and predictions may proceed.
projectStageDescription
string¦null
true
A description of the current stage of the project.
projectStatusError
string¦null
false
Project status error message.
rowCount
integer¦null
true
Count of rows in project's dataset.
rowPercentage
number¦null
true
Percentage of rows in segment project's dataset comparing to original dataset.
Constraints on valid values for this parameter. Note that any of these fields may be omitted but at least one will always be present. The presence of a field indicates that the parameter in question will accept values in the corresponding format.
Indicates that the value can contain free-form ASCII text. If present, is an empty object. Note that ascii fields must be valid ASCII-encoded strings. Additionally, they may not contain semicolons or newlines.
Numeric constraints on a value of an array of floating-point numbers. If present, indicates that this parameter's value may be a JSON array of numbers (integer or floating point).
Numeric constraints on a value of an array of floating-point numbers. If present, indicates that this parameter's value may be a JSON array of integers.
Indicates that the value can contain free-form ASCII text. If present, is an empty object. Note that ascii fields must be valid ASCII-encoded strings. Additionally, they may not contain semicolons or newlines.
A dictionary cvScores with sub-dictionary keyed by partition_id, each partition_id is itself a dictionary keyed by metric_name where the value is the reading for that particular metric for the partition_id.
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
effectiveFeatureDerivationWindowEnd
integer
true
maximum: 0
Only available for time series projects. How many timeUnits into the past relative to the forecast point the feature derivation window should end.
effectiveFeatureDerivationWindowStart
integer
true
maximum: 0 (exclusive)
Only available for time series projects. How many timeUnits into the past relative to the forecast point the user needs to provide history for at prediction time. This can differ from the featureDerivationWindowStart set on the project due to the differencing method and period selected.
featurelistId
string¦null
true
The ID of the feature list used by the model.
featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
forecastWindowEnd
integer
true
minimum: 0
Only available for time series projects. How many timeUnits into the future relative to the forecast point the forecast window should end.
forecastWindowStart
integer
true
minimum: 0
Only available for time series projects. How many timeUnits into the future relative to the forecast point the forecast window should start.
hasFinetuners
boolean
false
Whether a model has fine tuners.
holdoutScore
number¦null
true
the holdout score of the model according to the project metric, if the score is available and the holdout is unlocked
holdoutStatus
string
true
the status of the holdout fold
id
string
true
The ID of the model.
isAugmented
boolean
false
Whether a model was trained using augmentation.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isNClustersDynamicallyDetermined
boolean
false
Whether number of clusters is dynamically determined. Only valid in unsupervised clustering projects.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
The link function the final modeler uses in the blueprint. If no link function exists, returns null.
metrics
object
true
Object where each metric has validation, backtesting, backtestingScores and holdout scores reported, or null if they have not been computed. The validation score will be the score of the first backtest, which will be computed during initial model training. The backtesting and backtestingScores scores are computed when requested via POST /api/v2/projects/{projectId}/datetimeModels/{modelId}/backtests/. The backtesting score is the average score across all backtests. The backtestingScores is an array of scores for each backtest, with the scores reported as null if the backtest score is unavailable. The holdout score is the score against the holdout data, using the training data defined in trainingInfo.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
the family model belongs to, e.g. SVM, GBM, etc.
modelFamilyFullName
string
true
The full name of the family that the model belongs to. For e.g., Support Vector Machine, Gradient Boosting Machine, etc.
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
monotonicDecreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
monotonicIncreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
nClusters
integer¦null
false
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
predictionThreshold
number
true
maximum: 1 minimum: 0
threshold used for binary classification in predictions.
predictionThresholdReadOnly
boolean
true
indicates whether modification of a predictions threshold is forbidden. Since v2.22 threshold modification is allowed.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
samplePct
string
true
always null for datetime models
samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
supportsComposableMl
boolean
true
indicates whether this model is supported in Composable ML.
supportsMonotonicConstraints
boolean
true
whether this model supports enforcing monotonic constraints
timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
featurelistId
string¦null
true
The ID of the feature list used by the model.
featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
id
string
true
The ID of the model.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
isTrainedOnGpu
boolean
true
Whether the model was trained using GPU workers.
metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
The full name of the family that the model belongs to (e.g., Support Vector Machine, Gradient Boosting Machine, etc.).
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
trainingRowCount
integer¦null
true
minimum: 0 (exclusive)
The number of rows used to train the model.
trainingStartDate
string(date-time)¦null
true
the start date of the dates in the partition column for the data used to train the model
Any extended message to include about the result. For example, if a job is submitted that is a duplicate of a job that has already been added to the queue, the message will mention that no new job was created.
The complexity score for this solution. Complexity score is a function of the mathematical operators used in the current solution. The complexity calculation can be tuned via model hyperparameters.
error
number¦null
true
The error for the current solution, as computed by eureqa using the errorMetric error metric. None if Eureqa model refitted existing solutions.
errorMetric
string
true
The Eureqa error metric identifier used to compute error metrics for this search. Note that Eureqa error metrics do not correspond 1:1 with DataRobot error metrics - the available metrics are not the same, and even equivalent metrics may be computed slightly differently.
eureqaSolutionId
string
true
The ID of the solution.
expression
string
true
The eureqa "solution string". This is a mathematical expression; human-readable but with strict syntax specifications defined by Eureqa.
expressionAnnotated
string
true
The expression, rendered with additional tags to assist in automatic parsing.
threshold
number
true
Classifier threshold selected by the backend, used to determine which model values are binned as positive and which are binned as negative. Must have a value between the binStart of the first bin and binEnd of the last bin.
The ID of the model to clone from. If omitted, will automatically search for and find the first leaderboard model created by the blueprint run that also created the solution associated with solutionId.
The complexity score for this solution. Complexity score is a function of the mathematical operators used in the current solution. The complexity calculation can be tuned via model hyperparameters.
error
number¦null
true
The error for the current solution, as computed by eureqa using the errorMetric error metric. None if Eureqa model refitted existing solutions.
errorMetric
string
true
The Eureqa error metric identifier used to compute error metrics for this search. Note that Eureqa error metrics do not correspond 1:1 with DataRobot error metrics - the available metrics are not the same, and even equivalent metrics may be computed slightly differently.
eureqaSolutionId
string
true
The ID of the solution.
expression
string
true
The eureqa "solution string". This is a mathematical expression; human-readable but with strict syntax specifications defined by Eureqa.
expressionAnnotated
string
true
The expression, rendered with additional tags to assist in automatic parsing.
A list which contains feature impact scores for each feature used by a model. If the model has more than 1000 features, the most important 1000 features are returned.
next
string(uri)¦null
true
URL for the next page of results or null.
previous
string(uri)¦null
true
URL for the next page of results or null.
ranRedundancyDetection
boolean
true
Indicates whether redundant feature identification was run while calculating this feature impact.
rowCount
integer¦null
true
The number of rows that was used to calculate feature impact. For the feature impact calculated with the default logic, without specifying the rowCount, we return null here.
shapBased
boolean
true
Indicates whether feature impact was calculated using Shapley values. True for anomaly detection models when the project is unsupervised, as permutation approach is not applicable. Note that supervised projects must use an alternative route for SHAP impact: /api/v2/projects/(projectId)/models/(modelId)/shapImpact/
Numeric constraints on a value of an array of floating-point numbers. If present, indicates that this parameter's value may be a JSON array of numbers (integer or floating point).
the ID of an existing model to use as a source of training parameters.
nClusters
integer
false
maximum: 100 minimum: 2
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
samplePct
number
false
the percentage of the dataset to use with the model. Only one of samplePct and trainingRowCount should be specified. The specified percentage should be between 0.0 and 100.0.
trainingRowCount
integer
false
the integer number of rows of the dataset to use with the model. Only one of samplePct and trainingRowCount should be specified.
Numeric constraints on a value of an array of floating-point numbers. If present, indicates that this parameter's value may be a JSON array of integers.
True if the model supports export of model's source code or compiled Java executable.
supportsCoefficients
boolean
true
True if model coefficients are available.
supportsConfusionMatrix
boolean
true
True if Confusion Matrix can be generated.
supportsDocumentTextExtractionSampleInsight
boolean
true
True if the model has document column(s) and document text extraction samples can be generated.
supportsEarlyStopping
boolean
false
True if this is an early stopping tree-based model and number of trained iterations can be retrieved.
supportsForecastAccuracy
boolean
true
True if Forecast Accuracy plots can be generated.
supportsForecastVsActual
boolean
true
True if Forecast vs Actual plots can be generated.
supportsImageActivationMaps
boolean
true
True if the model has image column(s) and activation maps can be generated.
supportsImageEmbedding
boolean
true
True if the model has image column(s) and image embeddings can be generated.
supportsLiftChart
boolean
true
True if Lift Chart can be generated.
supportsModelPackageExport
boolean
false
True if the model can be exported as a model package.
supportsModelTrainingMetrics
boolean
true
When True , the model will track and save key training metrics in an effort to communicate model accuracy throughout training, rather than at training completion.
supportsMonotonicConstraints
boolean
true
True if the model supports monotonic constraints.
supportsNNVisualizations
boolean
true
True if the model supports neuralNetworkVisualizations.
supportsPeriodAccuracy
boolean
false
True if Period Accuracy insights can be generated.
supportsPredictionExplanations
boolean
true
True if the model supports Prediction Explanations.
supportsPredictionIntervals
boolean
true
True if Prediction Intervals can be computed for predictions generated by this model.
supportsResiduals
boolean
true
When True, the model supports residuals and residuals data can be retrieved.
supportsRocCurve
boolean
true
True if ROC Curve can be generated.
supportsSeriesInsights
boolean
true
True if Series Insights can be generated.
supportsShap
boolean
true
True if the model supports Shapley package. i.e. Shapley based feature Importance
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
featurelistId
string¦null
true
The ID of the feature list used by the model.
featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
hasFinetuners
boolean
false
Whether a model has fine tuners.
id
string
true
The ID of the model.
isAugmented
boolean
false
Whether a model was trained using augmentation.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isNClustersDynamicallyDetermined
boolean
false
Whether number of clusters is dynamically determined. Only valid in unsupervised clustering projects.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
The link function the final modeler uses in the blueprint. If no link function exists, returns null.
metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
the family model belongs to, e.g. SVM, GBM, etc.
modelFamilyFullName
string
true
The full name of the family that the model belongs to. For e.g., Support Vector Machine, Gradient Boosting Machine, etc.
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
monotonicDecreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
monotonicIncreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
nClusters
integer¦null
false
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
predictionThreshold
number
true
maximum: 1 minimum: 0
threshold used for binary classification in predictions.
predictionThresholdReadOnly
boolean
true
indicates whether modification of a predictions threshold is forbidden. Since v2.22 threshold modification is allowed.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
supportsComposableMl
boolean
true
indicates whether this model is supported in Composable ML.
supportsMonotonicConstraints
boolean
true
whether this model supports enforcing monotonic constraints
timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
trainingRowCount
integer¦null
true
minimum: 0 (exclusive)
The number of rows used to train the model.
trainingStartDate
string(date-time)¦null
true
the start date of the dates in the partition column for the data used to train the model
(Deprecated in version v2.11) Renamed to knownInAdvanceFeatureNames. This parameter always has the same value as knownInAdvanceFeatureNames and will be removed in a future release.
featureNames
[string]
true
An array of the names of all features used by the specified model.
knownInAdvanceFeatureNames
[string]
true
An array of the names of time series known-in-advance features used by the specified model.
Whether inferring actual values from time series history data and automatically feeding them back for accuracy estimation is supported by this model package.
supportsChallengerModels
boolean
true
Whether Challenger Models are supported by this model package.
supportsFeatureDriftTracking
boolean
true
Whether Feature Drift is supported by this model package.
supportsHumilityRecommendedRules
boolean
true
Whether calculating values for recommended Humility Rules is supported by this model package.
supportsHumilityRules
boolean
true
Whether Humility Rules are supported by this model package.
supportsHumilityRulesDefaultCalculations
boolean
true
Whether calculating default values for Humility Rules is supported by this model package.
supportsPredictionWarning
boolean
true
Whether Prediction Warnings are supported by this model package.
supportsRetraining
boolean
false
Whether deployment supports retraining.
supportsScoringCodeDownload
boolean
false
Whether scoring code download is supported by this model package.
supportsSecondaryDatasets
boolean
true
If the deployments supports secondary datasets.
supportsSegmentedAnalysisDriftAndAccuracy
boolean
true
Whether tracking features in training and predictions data for segmented analysis is supported by this model package.
supportsShapBasedPredictionExplanations
boolean
true
Whether shap-based prediction explanations are supported by this model package.
supportsTargetDriftTracking
boolean
true
Whether Target Drift is supported by this model package.
Names of categorical features by which the training baseline was segmented. This allows for deployment prediction requests to be segmented by those same features. Segmenting the training baseline by these features allows for users to perform segmented analysis of Data Drift and Accuracy, and to compare the same subset of training and scoring data based on the selected segment attribute and segment value.
datasetName
string¦null
true
Name of dataset used to train the model
holdoutDataCatalogId
string¦null
true
ID for holdout data (returned from uploading a data set)
holdoutDataCatalogVersionId
string¦null
true
Version ID for holdout data (returned from uploading a data set)
holdoutDataCreatedAt
string¦null
false
Time when the holdout data item was created
holdoutDataCreatorEmail
string¦null
false
Email of the user who created the holdout data item
holdoutDataCreatorId
string¦null
false
ID of the creator of the holdout data item
holdoutDataCreatorName
string¦null
false
Name of the user who created the holdout data item
holdoutDatasetName
string¦null
true
Name of dataset used for model holdout
targetHistogramBaseline
string
false
Values used to establish the training baseline
trainingDataCatalogId
string¦null
true
ID for training data (returned from uploading a data set)
trainingDataCatalogVersionId
string¦null
true
Version ID for training data (returned from uploading a data set)
trainingDataCreatedAt
string¦null
false
Time when the training data item was created
trainingDataCreatorEmail
string¦null
false
Email of the user who created the training data item
trainingDataCreatorId
string¦null
false
ID of the creator of the training data item
trainingDataCreatorName
string¦null
false
Name of the user who created the training data item
trainingDataSize
integer
false
Number of rows in training data (used by DR models)
model description information for the model package
modelExecutionType
string
true
Type of model package. dedicated (native DataRobot models) and custom_inference_model (user added inference models) both execute on DataRobot prediction servers, external do not
Class names for prediction results. When target type is Binary, two class names are returned. The first element is the minority (positive) class and the second element is the majority (negative) class. Limited to 100 returned for Multiclass.
name
string
true
name of the target column
predictionProbabilitiesColumn
string¦null
true
Field or column name containing prediction probabilities
predictionThreshold
number¦null
true
maximum: 1 minimum: 0
Prediction threshold used for binary classification models
Date format for forecast date and forecast point column
datetimeColumnName
string¦null
true
Name of the forecast date column
effectiveFeatureDerivationWindowEnd
integer¦null
true
maximum: 0
Same concept as featureDerivationWindowEnd which is chosen by the user and based on the initial sampled data from the eda sample. When the dataset goes through aim, the pipeline reads the full dataset and figures out the "real" FDW (i.e., the effective FDW). For most models, eFDW is approximately the same as the FDW.
effectiveFeatureDerivationWindowStart
integer¦null
true
maximum: 0
Same concept as featureDerivationWindowStart which is chosen by the user and based on the initial sampled data from the eda sample. When the dataset goes through aim, the pipeline reads the full dataset and figures out the "real" FDW (i.e., the effective FDW). For most models, eFDW is approximately the same as the FDW.
featureDerivationWindowEnd
integer¦null
true
maximum: 0
Negative number or zero defining how many time units of the forecast distances time unit into the past relative to the forecast point the feature derivation window should end.
featureDerivationWindowStart
integer¦null
true
maximum: 0
Negative number or zero defining how many time units of the forecast distances time unit into the past relative to the forecast point the feature derivation window should begin.
forecastDistanceColumnName
string¦null
true
Name of the forecast distance column
forecastDistances
[integer]¦null
true
List of integer forecast distances
forecastDistancesTimeUnit
string¦null
true
The time unit of forecast distances
forecastPointColumnName
string¦null
true
Name of the forecast point column
isCrossSeries
boolean¦null
true
true if the model is cross-series.
isNewSeriesSupport
boolean¦null
true
true if the model is optimized to support new series.
isTraditionalTimeSeries
boolean¦null
true
true if the model is traditional time series.
seriesColumnName
string¦null
true
Name of the series column in case of multi-series date
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
id
string
true
The ID of the model.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
isTrainedOnGpu
boolean
true
Whether the model was trained using GPU workers.
metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
The full name of the family that the model belongs to (e.g., Support Vector Machine, Gradient Boosting Machine, etc.).
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
any extended message to include about the result. For example, if a job is submitted that is a duplicate of a job that has already been added to the queue, the message will mention that no new job was created.
The Eureqa error metric identifier used to compute error metrics for this search. Note that Eureqa error metrics do not correspond 1:1 with DataRobot error metrics - the available metrics are not the same, and even equivalent metrics may be computed slightly differently.
The percentage of missing values in the training data
tasks
object
true
Information on individual tasks of the model which were used to process the feature. The names of properties will be task ids (which correspond to the ids used in the blueprint chart endpoints like GET /api/v2/projects/{projectId}/blueprints/{blueprintId}/blueprintChart/) The corresponding value for each task will be of the form task described.
Human readable aggregated information about how the task handles missing values. The following descriptions may be present: what value is imputed for missing values, whether the feature being missing is treated as a feature by the task, whether missing values are treated as infrequent values, whether infrequent values are treated as missing values, and whether missing values are ignored.
name
string
true
Task name, e.g. 'Ordinal encoding of categorical variables'
Any extended message to include about the result. For example, if a job is submitted that is a duplicate of a job that has already been added to the queue, the message will mention that no new job was created.
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
featurelistId
string¦null
true
The ID of the feature list used by the model.
featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
hasFinetuners
boolean
false
Whether a model has fine tuners.
id
string
true
The ID of the model.
isAugmented
boolean
false
Whether a model was trained using augmentation.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isNClustersDynamicallyDetermined
boolean
false
Whether number of clusters is dynamically determined. Only valid in unsupervised clustering projects.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
The link function the final modeler uses in the blueprint. If no link function exists, returns null.
metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
the family model belongs to, e.g. SVM, GBM, etc.
modelFamilyFullName
string
true
The full name of the family that the model belongs to. For e.g., Support Vector Machine, Gradient Boosting Machine, etc.
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
monotonicDecreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
monotonicIncreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
nClusters
integer¦null
false
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
predictionThreshold
number
true
maximum: 1 minimum: 0
threshold used for binary classification in predictions.
predictionThresholdReadOnly
boolean
true
indicates whether modification of a predictions threshold is forbidden. Since v2.22 threshold modification is allowed.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
ruleCount
integer
true
the number of rules used to create this model
rulesetId
integer
true
the ID of the ruleset this model uses
samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
score
number
true
the validation score of the models ruleset
supportsComposableMl
boolean
true
indicates whether this model is supported in Composable ML.
supportsMonotonicConstraints
boolean
true
whether this model supports enforcing monotonic constraints
timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
trainingRowCount
integer¦null
true
minimum: 0 (exclusive)
The number of rows used to train the model.
trainingStartDate
string(date-time)¦null
true
the start date of the dates in the partition column for the data used to train the model
The name that was used for the rating table. May differ from the ratingTableName in the request, as names are trimmed and a suffix added to ensure all rating tables derived from the same model have unique names
the actual records. Each element of the array has the same schema as if retrieving the table individually from GET /api/v2/projects/(projectId)/ratingTables/(ratingTableId)/
next
string(uri)¦null
true
URL pointing to the next page (if null, there is no next page)
previous
string(uri)¦null
true
URL pointing to the previous page (if null, there is no previous page)
Identifies which setting defines the training size of the model when making predictions and scoring. Only used by datetime models.
featurelistId
string¦null
true
The ID of the feature list used by the model.
featurelistName
string¦null
true
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
hasFinetuners
boolean
false
Whether a model has fine tuners.
id
string
true
The ID of the model.
isAugmented
boolean
false
Whether a model was trained using augmentation.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isNClustersDynamicallyDetermined
boolean
false
Whether number of clusters is dynamically determined. Only valid in unsupervised clustering projects.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
The link function the final modeler uses in the blueprint. If no link function exists, returns null.
metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
the family model belongs to, e.g. SVM, GBM, etc.
modelFamilyFullName
string
true
The full name of the family that the model belongs to. For e.g., Support Vector Machine, Gradient Boosting Machine, etc.
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
monotonicDecreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
monotonicIncreasingFeaturelistId
string¦null
true
the ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
nClusters
integer¦null
false
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
predictionThreshold
number
true
maximum: 1 minimum: 0
threshold used for binary classification in predictions.
predictionThresholdReadOnly
boolean
true
indicates whether modification of a predictions threshold is forbidden. Since v2.22 threshold modification is allowed.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
ratingTableId
string
true
The rating table ID
samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
samplingMethod
string
false
indicates sampling method used to select training data in datetime models. For row-based project this is the way how requested number of rows are selected.For other projects (duration-based, start/end, project settings) - how specified percent of rows (timeWindowSamplePct) is selected from specified time window.
supportsComposableMl
boolean
true
indicates whether this model is supported in Composable ML.
supportsMonotonicConstraints
boolean
true
whether this model supports enforcing monotonic constraints
timeWindowSamplePct
integer¦null
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99, indicating the percentage of sampling within the time window. The points kept are determined by samplingMethod option. Will be null if no sampling was specified. Only used by datetime models.
trainingDuration
string¦null
true
the duration spanned by the dates in the partition column for the data used to train the model
trainingEndDate
string(date-time)¦null
true
the end date of the dates in the partition column for the data used to train the model
trainingRowCount
integer¦null
true
minimum: 0 (exclusive)
The number of rows used to train the model.
trainingStartDate
string(date-time)¦null
true
the start date of the dates in the partition column for the data used to train the model
If present, the reason why Accuracy Over Time plots cannot be generated for the model.
supportsAnomalyAssessment
string
false
If present, the reason why the Anomaly Assessment insight cannot be generated for the model.
supportsAnomalyOverTime
string
false
If present, the reason why Anomaly Over Time plots cannot be generated for the model.
supportsClusterInsights
string
false
If present, the reason why Cluster Insights cannot be generated for the model.
supportsConfusionMatrix
string
false
If present, the reason why Confusion Matrix cannot be generated for the model. There are some cases where Confusion Matrix is available but it was calculated using stacked predictions or in-sample predictions.
supportsDocumentTextExtractionSampleInsight
string
false
If present, the reason document text extraction sample insights are not supported for the model.
supportsForecastAccuracy
string
false
If present, the reason why Forecast Accuracy plots cannot be generated for the model.
supportsForecastVsActual
string
false
If present, the reason why Forecast vs Actual plots cannot be generated for the model.
supportsImageActivationMaps
string
false
If present, the reason image activation maps are not supported for the model.
supportsImageEmbedding
string
false
If present, the reason image embeddings are not supported for the model.
supportsLiftChart
string
false
If present, the reason why Lift Chart cannot be generated for the model. There are some cases where Lift Chart is available but it was calculated using stacked predictions or in-sample predictions.
supportsPeriodAccuracy
string
false
If present, the reason why Period Accuracy insights cannot be generated for the model.
supportsPredictionExplanations
string
false
If present, the reason why Prediction Explanations cannot be computed for the model.
supportsPredictionIntervals
string
false
If present, the reason why Prediction Intervals cannot be computed for the model.
supportsResiduals
string
false
If present, the reason why residuals are not available for the model. There are some cases where Residuals are available but they were calculated using stacked predictions or in-sample predictions.
supportsRocCurve
string
false
If present, the reason why ROC Curve cannot be generated for the model. There are some cases where ROC Curve is available but it was calculated using stacked predictions or in-sample predictions.
supportsSeriesInsights
string
false
If present, the reason why Series Insights cannot be generated for the model.
supportsStability
string
false
If present, the reason why Stability plots cannot be generated for the model.
Information on the creator of the registered model
description
string¦null
false
Description of the registered model
id
string
true
ID of the registered model
isArchived
boolean
true
Whether the model is archived
isGlobal
boolean
false
Whether the registered model is global (accessible to all users in the organization) or local(accessible only to the owner and the users with whom it has been explicitly shared)
lastVersionNum
integer
true
Latest version associated to this registered model
Make registered model global (accessible to all users in the organization) or local(accessible only to the owner and the users with whom it has been explicitly shared)
If specified, the new model will be trained using this featurelist. Otherwise, the model will be trained on the same feature list as the source model.
modelId
string
true
The ID of an existing model to use as the source for the training parameters.
monotonicDecreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
monotonicIncreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
nClusters
integer
false
maximum: 100 minimum: 2
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
samplingMethod
string
false
Defines how training data is selected if subsampling is used (e.g., if timeWindowSamplePct is specified). Can be either random or latest. If omitted, defaults to latest if trainingRowCount is used and random for other cases (e.g., if trainingDuration or useProjectSettings is specified). May only be specified for OTV projects.
timeWindowSamplePct
integer
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99 indicating the percentage of sampling within the time window. The points kept are determined by the value provided for the samplingMethod option. If specified, trainingRowCount may not be specified, and the specified model must either be a duration or selectedDateRange model, or one of trainingDuration or trainingStartDate and trainingEndDate must be specified.
trainingDuration
string(duration)
false
A duration string representing the training duration to use for training the new model. If specified, the model will be trained using the specified training duration. Otherwise, the original model's duration will be used. Only one of trainingRowCount, trainingDuration, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
trainingEndDate
string(date-time)
false
A datetime string representing the end date of the data to use for training this model. Note that only one of trainingDuration or trainingRowCount or trainingStartDate and trainingEndDate should be specified. If trainingStartDate and trainingEndDate are specified, the source model must be frozen.
trainingRowCount
integer
false
minimum: 0 (exclusive)
The number of rows of data that should be used to train the model. If not specified, the original model's row count will be used. Only one of trainingRowCount, trainingDuration, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
trainingStartDate
string(date-time)
false
A datetime string representing the start date of the data to use for training this model. Note that only one of trainingDuration or trainingRowCount or trainingStartDate and trainingEndDate should be specified. If trainingStartDate and trainingEndDate are specified, the source model must be frozen.
useProjectSettings
boolean
false
If True, the model will be trained using the previously-specified custom backtest training settings. Only one of trainingRowCount, trainingDuration, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
If specified, the model will be trained using that featurelist, otherwise the model will be trained on the same feature list as before.
modelId
string
true
The model to be retrained
monotonicDecreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no such constraints are enforced.
monotonicIncreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no such constraints are enforced.
nClusters
integer
false
maximum: 100 minimum: 2
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
samplePct
number
false
The percentage of the dataset to use to use to train the model. The specified percentage should be between 0 and 100. If not specified, original model sample percent will be used.
scoringType
string
false
Validation is available for any partitioning. If the project uses cross validation, crossValidation may be used to indicate that all available training/validation combinations should be used.
trainingRowCount
integer
false
The number of rows to use to train the model. If not specified, original model training row count will be used.
The complexity score for this solution. Complexity score is a function of the mathematical operators used in the current solution. The complexity calculation can be tuned via model hyperparameters.
error
number¦null
true
The error for the current solution, as computed by eureqa using the errorMetric error metric. None if Eureqa model refitted existing solutions.
eureqaSolutionId
string
true
The ID of the solution.
expression
string
true
The eureqa "solution string". This is a mathematical expression; human-readable but with strict syntax specifications defined by Eureqa.
expressionAnnotated
string
true
The expression, rendered with additional tags to assist in automatic parsing.
The ID of an existing model to use as the source for the training parameters.
samplingMethod
string
false
Defines how training data is selected if subsampling is used (e.g., if timeWindowSamplePct is specified). Can be either random or latest. If omitted, defaults to latest if trainingRowCount is used and random for other cases (e.g., if trainingDuration or useProjectSettings is specified). May only be specified for OTV projects.
timeWindowSamplePct
integer
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99 indicating the percentage of sampling within the time window. The points kept are determined by the value provided for the samplingMethod option. If specified, trainingRowCount may not be specified, and the specified model must either be a duration or selectedDateRange model, or one of trainingDuration or trainingStartDate and trainingEndDate must be specified.
trainingDuration
string(duration)
false
A duration string representing the training duration for the submitted model. Only one of trainingDuration, trainingRowCount, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
trainingEndDate
string(date-time)
false
A datetime string representing the end date of the data to use for training this model. If specified, trainingStartDate must also be specified. Only one of trainingDuration, trainingRowCount, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
trainingRowCount
integer
false
minimum: 0 (exclusive)
The number of rows of data that should be used when training this model. Only one of trainingDuration, trainingRowCount, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
trainingStartDate
string(date-time)
false
A datetime string representing the start date of the data to use for training this model. If specified, trainingEndDate must also be specified. Only one of trainingDuration, trainingRowCount, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
useProjectSettings
boolean
false
If True, the model will be trained using the previously-specified custom backtest training settings. Only one of trainingDuration, trainingRowCount, trainingStartDate and trainingEndDate, or useProjectSettings may be specified.
If specified, the model will be trained using this featurelist. If not specified, the recommended featurelist for the specified blueprint will be used. If there is no recommended featurelist, the project's default will be used.
monotonicDecreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no constraints will be enforced. If omitted, the project default is used. May only be specified for OTV projects.
monotonicIncreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no constraints will be enforced. If omitted, the project default is used. May only be specified for OTV projects.
nClusters
integer
false
maximum: 100 minimum: 2
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
samplingMethod
string
false
Defines how training data is selected if subsampling is used (e.g., if timeWindowSamplePct is specified). Can be either random or latest. If omitted, defaults to latest if trainingRowCount is used and random for other cases (e.g., if trainingDuration or useProjectSettings is specified). May only be specified for OTV projects.
sourceProjectId
string
false
The project the blueprint comes from. Required only if the blueprintId comes from a different project.
timeWindowSamplePct
integer
false
maximum: 100 (exclusive) minimum: 0 (exclusive)
An integer between 1 and 99 indicating the percentage of sampling within the time window. The points kept are determined by the value provided for the samplingMethod option. If specified, trainingRowCount may not be specified, and the specified model must either be a duration or selectedDateRange model, or one of trainingDuration or trainingStartDate and trainingEndDate must be specified.
trainingDuration
string(duration)
false
A duration string representing the training duration for the submitted model.
trainingRowCount
integer
false
minimum: 0 (exclusive)
The number of rows of data that should be used when training this model.
useProjectSettings
boolean
false
If True, the model will be trained using the previously-specified custom backtest training settings.
If specified, the model will be trained using this featurelist. If not specified, the recommended featurelist for the specified blueprint will be used. If there is no recommended featurelist, the project's default will be used.
monotonicDecreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically decreasing relationship to the target. If null, no constraints will be enforced. If omitted, the project default is used.
monotonicIncreasingFeaturelistId
string¦null
false
The ID of the featurelist that defines the set of features with a monotonically increasing relationship to the target. If null, no constraints will be enforced. If omitted, the project default is used.
nClusters
integer
false
maximum: 100 minimum: 2
The number of clusters to use in the specified unsupervised clustering model. Only valid in unsupervised clustering projects.
samplePct
number
false
maximum: 100 minimum: 0 (exclusive)
The percentage of the dataset to use with the model. Only one of samplePct and trainingRowCount should be specified. The specified percentage should be between 0 and 100.
scoringType
string
false
Validation is available for any partitioning. If the project uses cross validation, crossValidation may be used to indicate that all available training/validation combinations should be used.
sourceProjectId
string
false
The project the blueprint comes from. Required only if the blueprintId comes from a different project.
trainingRowCount
integer
false
An integer representing the number of rows of the dataset to use with the model. Only one of samplePct and trainingRowCount should be specified.
Constraints on valid values for this parameter. Note that any of these fields may be omitted but at least one will always be present. The presence of a field indicates that the parameter in question will accept values in the corresponding format.
currentValue
any
true
The single value or list of values of the parameter that were grid searched. Depending on the grid search specification, could be a single fixed value (no grid search), a list of discrete values, or a range.
oneOf
Name
Type
Required
Restrictions
Description
» anonymous
any
false
none
anyOf
Name
Type
Required
Restrictions
Description
»» anonymous
string
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
integer
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
boolean
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
number
false
none
xor
Name
Type
Required
Restrictions
Description
» anonymous
[anyOf]
false
none
anyOf
Name
Type
Required
Restrictions
Description
»» anonymous
string
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
integer
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
boolean
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
number
false
none
continued
Name
Type
Required
Restrictions
Description
defaultValue
any
true
The actual value used to train the model; either the single value of the parameter specified before training, or the best value from the list of grid-searched values (based on current_value).
oneOf
Name
Type
Required
Restrictions
Description
» anonymous
any
false
none
anyOf
Name
Type
Required
Restrictions
Description
»» anonymous
string
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
integer
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
boolean
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
number
false
none
xor
Name
Type
Required
Restrictions
Description
» anonymous
[anyOf]
false
none
anyOf
Name
Type
Required
Restrictions
Description
»» anonymous
string
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
integer
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
boolean
false
none
or
Name
Type
Required
Restrictions
Description
»» anonymous
number
false
none
continued
Name
Type
Required
Restrictions
Description
parameterId
string
true
Unique (per-blueprint) identifier of this parameter. This is the identifier used to specify which parameter to tune when make a new advanced tuning request.
parameterName
string
true
Name of the parameter.
taskName
string
true
Human-readable name of the task that this parameter belongs to.
The name of the feature list used by the model. If null, themodel was trained on multiple feature lists.
id
string
true
The ID of the model.
isFrozen
boolean
true
Indicates whether the model is frozen, i.e., uses tuning parameters from a parent model.
isStarred
boolean
true
Indicates whether the model has been starred.
isTrainedIntoHoldout
boolean
true
Indicates if model used holdout data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed validation size.
isTrainedIntoValidation
boolean
true
Indicates if model used validation data for training. This can happen for time-aware models using trainingStartDate/trainingEndDate parameters or when the model's training row count was greater than the max allowed training size.
isTrainedOnGpu
boolean
true
Whether the model was trained using GPU workers.
metrics
object
true
The performance of the model according to various metrics, where each metric has validation, crossValidation, holdout, and training scores reported, or null if they have not been computed.
modelCategory
string
true
Indicates the type of model. Returns prime for DataRobot Prime models, blend for blender models, combined for combined models, and model for all other models.
modelFamily
string
true
The full name of the family that the model belongs to (e.g., Support Vector Machine, Gradient Boosting Machine, etc.).
modelNumber
integer¦null
true
minimum: 0 (exclusive)
The model number from the Leaderboard.
modelType
string
true
Identifies the model (e.g.,Nystroem Kernel SVM Regressor).
numberOfClusters
integer¦null
true
The number of clusters in the unsupervised clustering model. Only present in unsupervised clustering projects.
parentModelId
string¦null
true
The ID of the parent model if the model is frozen or a result of incremental learning. Null otherwise.
processes
[string]¦null
true
maxItems: 100
The list of processes used by the model.
projectId
string
true
The ID of the project to which the model belongs.
samplePct
number¦null
true
minimum: 0 (exclusive)
The percentage of the dataset used in training the model.
Whether the org/group/user should be able to share with others.If true, the org/group/user will be able to grant any role up to and includingtheir own to other orgs/groups/user. If role is NO_ROLEcanShare is ignored.
role
string¦null
true
The role to set on the entity. When it is None, the role of this user will be removedfrom this entity.
username
string
true
Username of the user to update the access role for.