Get Predictive Model Measures
  • 18 Apr 2025
  • 1 Minute to read
  • Dark
    Light

Get Predictive Model Measures

  • Dark
    Light

Article summary

Step Details
Introduced in VersionProcess Mining 3.2
Last Modified in VersionProcess Mining 3.2
LocationProcess Mining  > Predictions



The Get Predictive Model Measures step in the Decisions Flow toolbox enables users to retrieve a predictive model's performance metrics from a Decisions instance. This provides an alternative to manually checking results in the Process Mining > Prediction > Prediction Center dashboard.

When a Predictive Model ID is provided as input, the step returns a structured set of evaluation metrics describing the model's test data performance. These metrics are critical for assessing the model's ability to predict outcomes and comparing different models or configurations.


Properties

Inputs

PropertyDescriptionData Type
Predictive Model ID
The ID of the model available in the Process Mining instance.
Int64

Outputs

PropertyDescriptionData Type
GetPredictiveModelMeasures_OutputA set of model evaluation metrics:
  • Accuracy: Proportion of total predictions that are correct.
  • Precision: Proportion of positive identifications that were actually correct.
  • Recall: Proportion of actual positives correctly identified.
  • F-Score: Harmonic mean of precision and recall.
  • AUC: Area under the ROC curve.
  • True Positive: Number of correctly predicted positive cases.
  • True Negative: Number of correctly predicted negative cases.
  • False Positive: Number of negative cases incorrectly predicted as positive.
  • False Negative: Number of positive cases incorrectly predicted as negative.
--

Example

This example illustrates how to obtain and use the prediction model ID in the Decisions workflow to get the model's evaluation metrics:

  1. Log in to the Process Mining dashboard.
  2. Navigate to Prediction > Prediction Center dashboard to obtain the model ID. Also, check the status, which currently indicates that the model is Ready. 
  3. Log in to the Decisions instance and create a Flow.
  4. Drag and drop the Get Predictive Model Measures Step from Toolbox > All Steps > Process Mining > Prediction.
  5. Select the Get Predictive Model Measures Step and open the Properties on the top right corner.
  6. Enter the model ID obtained from the Process Mining > Prediction > Prediction Center dashboard in the Predictive Model ID input field.
  7. Click on Debug Flow to test the Step and see the model's evaluation metrics.

Was this article helpful?