@KshitijP On the Predictors tab of an adaptive model rule, you can delete any field you want. However, it is highly recommended to have as many - uncorrelated - predictors as possible and let the models figure out which ones to use. The impact of inactive predictors on system performance is minimal. Moreover, fields that are currently inactive may at some point in time become relevant, deleting them would block that opportunity. In the Data Scientist mission on Pega Academy, more information on configuring adaptive models is available.
@thijh : Thanks a lot for your quick response and suggestions. Actually the backend data for the predictors has been stopped refreshing and no plan to get it back again hence we are planning to remove it first from the Models before we plan to remove the properties from database mapping.
Quick question : Is my understanding is correct removing these predictors will not impact the previous learning of the model(technically will not reset the model to start learning from zero) ?
Is there any guidelines/steps that we should follow before deleting them from models?
Removing predictors does not reset the models. Removing active predictors will of course have some (possibly temporary) effect on the models, but as you intend to only remove inactive predictors, this will not be the case.
Please be aware that removal of a predictor is at the rule level, so it impacts all the “instances” of the model.