In this project I will build and analyze different machine-learning models to predict credit risk.
In the Resampling_Credit_Risk notebook I will perform the following tests and compare the results:
-
Oversample data using Naive Random Oversampler / SMOTE algorithms.
-
Undersample data using Cluster Centroids algorithm.
-
Over / undersample using a combination SMOTEENN algorithm.
In the Ensemble_Learning_Credit_Risk notebook I will run two different ensemble classifiers (Balanced Random Forest / Easy Ensemble) and compare the results.