mlToolbox

Python functions for manipulating pandas dataframes, useful for behavioral analyses, general data exploration, and modeling.

Dependencies

• numpy
• scipy
• pandas
• scikit-learn

pandas_helpers.py

Contains various functions to preprocess pandas dataframes (e.g., transforming features, imputing missing data), as well as functions to fit and evaluate various models using sklearn.

• fit_evaluate_models: Given samples (X) and a continuous or categorical vector to predict (y), train/test various models using a cross-validation approach, and return a pandas dataframe with evaluative metrics matching the type of dependent variable (e.g., coefficient of determination for a continuous DV, accuracy for a categorical DV). Optionally scale the features (across samples), or add in polynomial + interaction features.