Skip to content

On-premise users: click in-app to access the full platform documentation for your version of DataRobot.

DataRobot API resources > API reference documentation > Batch Prediction API > Output formats

Output format

DataRobot returns predictions in a columnar table format. Each example value is followed by the data type it belongs to. The columns returned are determined by model type, as described below.

Note

DataRobot allows prediction output to many different databases that all have unique versions of a string (e.g., some may call it TEXT while others may call it VARCHAR). As a result, DataRobot cannot provide implementation-specific data types.

Regression models

Prediction label
Column name <target_name>_PREDICTION
Data type Numeric
Example name revenue_PREDICTION
Example value 493822.12
Description The predicted value.

Binary classification models

Positive label
Column name <target_name>_<positive_label>_PREDICTION
Data type Numeric
Example name isbadbuy_1_PREDICTION
Example value 0.28
Description The float probability of the positive label.
Negative label
Column name <target_name>_<negative_label>_PREDICTION
Data type Numeric
Example name isbadbuy_0_PREDICTION
Example value 0.72
Description The float probability of the negative label.
Prediction label
Column name <target_name>_PREDICTION
Data type Text
Example name isbadbuy_PREDICTION
Example value 0
Description The predicted label of the classification.
Threshold label
Column name THRESHOLD
Data type Numeric
Example name THRESHOLD
Example value 0.5
Description The float prediction threshold used for determining the label.
Positive class label
Column name POSITIVE_CLASS
Data type Text
Example name POSITIVE_CLASS
Example value 1
Description The label configured as the positive class.

Multiclass classification models

Prediction label
Column name <target_name>_PREDICTION
Data type Text
Example name species_PREDICTION
Example value lion
Description The predicted label of the classification.
Prediction class label (for each class)
Column name <target_name>_<class_label>_PREDICTION
Data type Numeric
Description The float probability for each class.
Example classifications
Example name Example value
species_cat_PREDICTION 0.28
species_lion_PREDICTION 0.24
species_lynx_PREDICTION 0.48

Time series models

Note

These output columns are available for time series regression, classification, and anomaly detection models.

Time series model columns Description Data type
<SERIES_ID_COLUMN_NAME> Contains the series ID the row belongs to.

Functions as a passthrough column and returns the unaltered column name and values provided in the scoring data.		 Text
FORECAST_POINT Contains the forecast point timestamp.

Unless you request historical time series predictions, the output value is the same for all rows with the same forecast point (but different for each unique forecast distance).		 Date
<TIME_COLUMN_NAME> Contains the time series timestamp.

Functions as a passthrough column and returns the unaltered column name and values provided in the scoring data. (This returns the same value as the originalFormatTimestamp field returned by time series models.)		 Date
FORECAST_DISTANCE Contains the numeric forecast distance returned by time series models. Numeric

Prediction status

Prediction status label
Column name prediction_status
Data type Text
Description A row-by-row status containing either OK or a string error message describing why the prediction did not succeed.
Example value Could not convert date field to date format YYYY-MM-DD
Example value OK

Prediction warnings

If prediction warnings are enabled for your job, DataRobot returns an additional column.

Prediction warnings label
Column name IS_OUTLIER_PREDICTION
Data type Text
Description Whether the prediction is outside the calculated prediction boundaries.
Example values
Column Example value
Data type Text
IS_OUTLIER_PREDICTION True
IS_OUTLIER_PREDICTION False

Deployment approval status

If the approval workflow is enabled for your deployment, the output schema will contain an extra column showing the deployment approval status.

Deployment status label
Column name DEPLOYMENT_APPROVAL_STATUS
Data type Text/td>
Description Whether the deployment was approved.
Example value PENDING

Prediction Explanations

You can request Prediction Explanations be returned with your predictions by setting the maxExplanations job parameter to a non-zero value. You can also set thresholds for computing explanations. If you do not configure a threshold, DataRobot computes explanations for every row.

Prediction Explanation parameters
Job parameter Description Example value Data type
maxExplanations (Optional) Compute up to this number of explanations. 10 Integer
thresholdHigh (Optional) Limit explanations to predictions above this threshold. 0.5 Float
thresholdLow (Optional) Limit explanations to predictions below this threshold. 0.15 Float

If Prediction Explanations are requested, DataRobot returns four extra columns for each explanation in the format EXPLANATION_<n>_IDENTIFIER (where n is the feature explanation index, from 1 to the maximum number of explanations requested). The returned columns are:

Prediction Explanation columns
Column Description Data type
EXPLANATION__FEATURE_NAME The feature name this explanation covers. Text
EXPLANATION__STRENGTH The feature strength as a float. Numeric
EXPLANATION__QUALITATIVE_STRENGTH The feature strength as a string, a plus or minus indicator from +++ to ---. Text
EXPLANATION__ACTUAL_VALUE The feature associated with this explanation. Text

Prediction Explanation examples

Name Value
EXPLANATION_1_FEATURE_NAME loan_status
EXPLANATION_1_ACTUAL_VALUE Charged Off
EXPLANATION_1_STRENGTH 1.380291221709652
EXPLANATION_1_QUALITATIVE_STRENGTH +++
Name Value
EXPLANATION_1_FEATURE_NAME loan_status
EXPLANATION_1_ACTUAL_VALUE Fully Paid
EXPLANATION_1_STRENGTH -1.2145340858375335
EXPLANATION_1_QUALITATIVE_STRENGTH ---

Passthrough columns

Passthrough columns you request are passed verbatim. If they conflict with any of the above names, the job is rejected.

Association ID

If your deployment was configured with an association ID for accuracy, all result sets will have that column passed through from the source data automatically.

Output filters

Use the following job configuration properties to control whether to display only specific class probabilities or none at all.

Output filter parameters
Job parameter Description Example value Data type
includeProbabilities (Optional) Include probabilities for all classes; defaults to true. true Boolean
includeProbabilitiesClasses (Optional) Include only probabilities for classes listed in the given array; defaults to an empty array []. ['setosa', 'versicolor'] Boolean
includePredictionStatus (Optional) Include the prediction_status column in the output; defaults to false. true Boolean

Note

For binary classification, includeProbabilities also controls the THRESHOLD and POSITIVE_CLASS columns.

Column name remapping

If your use case has a strict output schema that does not match the DataRobot output, you can rename and remove any columns from the output using the columnNamesRemapping job configuration property.

Output column name remapping parameters
Job parameter Description Example value
columnNamesRemapping (Optional) Provide a list of items to remap (rename or remove columns from) the output from this job. Set an outputName for the column to null or false to ignore it. [{'inputName': 'isbadbuy_1_PREDICTION', 'outputName':'prediction'}, {'inputName': 'isbadbuy_0_PREDICTION', 'outputName': null}]
Updated September 6, 2024