feature_ratioparameters within setup. Feature interaction creates new features by multiplying two variables (a * b), while feature ratios create new features but by calculating the ratios of existing features (a / b).
polynomial_featuresparameter within setup.
polynomial_thresholdparameter which uses feature importance based on the combination of Random Forest, AdaBoost and Linear correlation to filter out the non important polynomial features.
polynomial_degreecan be used for defining number of degrees to be considered in feature creation.
group_featuresparameter within setup.
bin_numeric_featuresparameter within setup. PyCaret uses the ‘sturges’ rule to determine the number of bins and also uses K-Means clustering to convert continuous numeric features into categorical features.
combine_rare_levelsparameter within setup.
create_clustersparameter within setup.