Paper Title
Credit Risk Prediction Using Machine Learning Analytics: An Ensemble Model
Abstract
The occurrence of the likelihood that a borrower would payback his debt to his lender (bank) is quite a challenging
task. Various factors must be taken into consideration for deciding on advancing any type of loans. This decision making
process requires sufficient quantitative data for computation purpose which is a very time consuming and involves a risk
ofhuman errors. To overcome this tedious and hectic task, an expert system named Expert System for Credit Risk Prediction
usingExtra-Trees (ESCRPET) has been proposed that uses acombination of oversampling technique,„Synthetic
Minority Oversampling Technique‟ (SMOTE) and undersampling technique, „Edited Nearest Neighbor‟ (ENN) to deal with
class imbalance problem,uses Boruta feature selection techniqueto select relevant features that are useful for model training
and finally classification is done using Extra-Trees ensemble bagging technique. The performance of the proposed model is
measured in terms of accuracy and f1-score using 10-folds cross validation and is compared with single classifier based
models, ensemble models and various models present in literature.
Keywords - Credit Risk, SMOTE, ENN, Boruta, Extra Tree