Credit Card Score Prediction Using Machine Learning

The review of credit issuance decisions begins to enhance the decision-making processes of manual judgement and statistical analysis, which considerably improves the reliability and efficiency of credit issuance decisions as financial institution databases grow in size. As one of the most essential statistical tools, machine learning algorithms have grown in importance in assisting with credit approval decisions. However, prediction performance has differed among prediction models due to the varying algorithms of each model and the selection of corresponding parameters in a specific model. To improve model construction in the credit scoring process and better analyse the forecast effectiveness of prevalent models, Based on a predetermined performance objective, this work assesses the prediction accuracy of numerous regression models and classifiers and offers an ideal model with the maximum prediction accuracy. Artificial Neural Network (ANN), Decision tree, Random Forest. Classifier are among the experimental models used in the analysis Ann performed well in the investigated model, with the greatest performance score of Balanced.

Keywords: Credit card, Prediction, Artificial Neural Network (ANN), Decision tree, Random Forest.