Defaulters-Risk-Prediction-Banking

Buliding a predictive model to help the banks in predicting the probability of a customer defaulting in the future based on the historic data is the ultimate objective of the project . For which structured methodology is followed, in which the historic data of the customers are cleaned , splitted for testing and training , independency are checked ,classification algorithms were used like J48,Naïve Bayes, MLP,base line accuracy for the available data set is calculated, models were repeatedly tested and trained with Hold out and K-fold cross validation methods ,significant features are selected and the models were re evaluated for accuracy and precision comparison , model errors are address with the ROC .Individual model outputs are captured and compared with each other to get the best model for the Defaulter prediction.