Amex Credit Risk Analysis Challenge is conducted by American Express part of MS4610 Introduction to Data Analytics
Data has 47 transformed features for training the machine learning models. Creation of new features is avoided due to the tranformation of the original features. Describtion of features helps in creating new relavant features. The final model is the stacked ensemble model of CatBoost and Random Forest Classifier with weights 0.9 and 0.1 respectively.
The details of the data is availble in the problem Statement. Complete inoformation regarding the dataset and the model training process is shared in the Amex_project_report.pdf file. Jupyter notebook IITM_AMEX_project.ipynb gives the hands on implementation of the training discussed in the project report.