/Predicting-Bankruptcy-of-firms

In class Kaggle competition on predicting bankruptcy of a firm

Primary LanguageJupyter Notebook

Predicting Bankruptcy for firms

-Used K-Means for feature reduction

-Applied classification algorithms like Random Forest, XGBoost, Light GBM and Stacked Ensembles using evaluation metric as AUC to obtain the probability of each firm.

-Performed SMOTE to upsample the minority class

-Achieved an AUC score of 0.96 on test data