Metadata file properties for external models in machine learning services

objective

The objective of the model that you want to predict. Enter a meaningful value, for example, Churn.

outcomeType

The type of outcome that the model predicts. The following values are available:

BINARY

Set this value for binary models that predict one of two possible outcome categories, for example, Churn and Loyal.

CATEGORICAL

Set this value for categorical models that predict one of more than two possible outcome categories, for example, Red, Green, and Blue.

CONTINUOUS

Set this value for continuous models that predict the outcome between a minimum and a maximum value, for example, between 1 and 99.

expectedPerformanceMeasure

The metric by which you measure expected performance. The following values are available:

AUC

Shows the total predictive performance for binary models in the Area Under the Curve (AUC) measurement unit. Models with an AUC of 50 provide random outcomes, while models with an AUC of 100 predict the outcome perfectly.

F-score

Shows the weighted harmonic mean of precision and recall for categorical models, where precision is the number of correct positive results divided by the number of all positive results returned by the classifier, and recall is the number of correct positive results divided by the number of all relevant samples. An F-score of 1 means perfect precision and recall, while 0 means no precision or recall.

RMSE

Shows the root-mean-square error value for continuous models that is calculated as the square root of the average of squared errors. In this measure of predictive power, a number represents the difference between the predicted outcomes and the actual outcomes, where 0 means flawless performance.

expectedPerformance

Note: This is an optional property.

A numeric value that represents the expected predictive performance of the model. For AUC and F-score models, set a decimal value between 0 and 100. For RMSE models, set any decimal value.

framework

Note: For Google AI models, do not specify this property. The framework property is automatically fetched from the Google AI platform. For more information about the supported Google AI Platform models, see Supported Google AI Platform models.

The framework property determines the input format and output format of the model.

For Amazon SageMaker models, the following values are available:

• xgboost
• tensorflow
• kmeansclustering
• knn
• linearlearner
• randomcutforest

For more information about the supported input and output formats for Amazon SageMaker models, see Supported Amazon SageMaker models.

modelingTechnique

The modeling technique that determines how the model is created, for example, Random forest or XGBoost.

The transparency score is based on the modeling technique. For more information about model transparency, see the Model transparency for predictive models article on Pega Community.

outcomes

Use this property to specify the outcomes that the model predicts. The outcomes depend on the model type:

• For binary outcome models, enter two values that represent the possible outcomes. The first value is the outcome for which you want to predict the probability, and the second value is the alternative outcome.

  For example, to predict whether a customer is likely to accept an offer, specify the property as follows:

  "outcomes" : {
      "values": [ "Accept","Reject" ]
   }

• For categorical outcome models, enter more than two values that represent the possible outcomes.

  For example, to predict a call context, specify the property as follows:

  "outcomes" : {
      "values": [ "Complaint","Credit Limit","Customer Service","Other" ]
  }

• For continuous outcome models, enter minimum and maximum outcome values. The first value is the lowest possible outcome, and the second value is the highest possible outcome.

  For example, to predict a customer's credit rating on a scale from 300 to 850, specify the property as follows:

  "outcomes": {
      "range": [300, 850]
  }