Combinations_of_variables_model_evaluation Class Reference

Detailed Description

Builds, evaluates and ranks statistical models from a set of variables.

Given an estimator (linear regression, knn classifier, ...) and a dataset consisting of features or dependent variables and a response or dependent variable, this class does three things:

– 1) Generate all combinations of independent variables up to a certain size.

– 2) Reports the predictions obtained through multiple rounds of cross-validation.

– 3) Reports statistics about predictions for every combination of variables using a provided custom class.