This project aims at solvinf two problem statements. • To evaluate the loan credit risk. • To detect whether the insurance claim is fraudulent or not.
The data that I have used is Kaggle data. The dataset consisted records of 82,000 customers with 16 attributes and 1000 customers with 39 attributes, respectively.
To fit the data, the classification models that we wish to compare are Logistic Regression, KNN Classifier, Random Forest Classifier and Gradiant Boosting Classifier. I have split the entire data frame into 80% Training Set and 20% Test set, where I hold out the Test Set for final model evaluation. Using the models with best score I have further deployed the app.