How do we categorize Predictors and Potential Predictors in Prediction models
Hello Team,
For the Feature Finder Option in PegaInfinity23 version, In the Potential predictors tab we have potential predictors and predictors attributes.
How do we categorize predictors and Potential Predictors, Because even if the mutual information score is low, they are categorized as potential predictors and few attributes with higher mutual information score are categorized as predictors. can anyone explain based on what criteria pega is categorizing them.
Thanks and regards
Swetha.
Accepted Solution
Updated: 28 Feb 2024 7:37 EST
Pegasystems Inc.
NL
Hi @Swetha,
A "Predictor" is an attribute that is currently being used in the model to predict the outcome, while a "Potential Predictor" is an attribute that is not currently being used in the model. Any field that is not already a predictor and has a mutual information score greater than 0 can be tagged as Potential Predictor. It is up to the experts to combine the mutual information score (higher the better ofcourse) and domain knowledge to pick from potential predictors. You can check out the documentation for further information.
Thanks