Forecasting Customer Bank Behavior Using Weka and Classification Algorithms

Banks derive significant profits from loans, necessitating careful customer selection to mitigate default risks. This study analyzes customer behavior at Société Générale Algeria Bank using three classification algorithms and using Weka software. A dataset comprising 300 customers was evaluated with three algorithms: Bayesian Network, Naïve Bayes, and the J48 Decision Tree. The goal of this study was to identify the most effective algorithm for classifying prospective customers as acceptable or not for loan approval. The performance of these algorithms was compared, focusing on accuracy and reliability. The results of this study indicated that the J48 Decision Tree algorithm outperformed the other methods, demonstrating superior classification accuracy. This suggests its potential as a robust tool for optimizing decision-making processes in the bank’s loan system. By integrating the J48 Decision Tree into its operations, the bank could enhance its ability to identify suitable customers, minimize risks, and ensure sustainable profitability in its lending practices.