Customer Credit Risk: Application and Evaluation of Machine Learning and Deep Learning Models

Financial institutions rely on credit risk evaluation or credit scoring models for decision-making, as these models analyse the creditworthiness of customers. Machine learning models can predict with high accuracy, and they are widely utilized in this sector. This study selected a dataset from openml.org, and after data pre-processing, visualization, and exploratory data analysis, seven algorithms were applied. The models were evaluated using various performance metrics. Support Vector Machines (SVM) achieved the highest accuracy (80.67%) with a good Recall of 93.55%. Also, when dimensions were reduced to 10 (from 20) through Principal Component Analysis, SVM demonstrated the highest accuracy (83.57%), an F1 score of 84.70%, a Recall of 88.84%, and an ROC AUC Score of 83.44%. By generating synthetic records using SMOTE, the open-source algorithm Extreme Gradient Boosting achieved the highest accuracy score of 83.3% and an ROC AUC of 83.29%. Future work may involve tuning the hyperparameters of these algorithms to improve other performance metrics.