Financial Prediction Models in Banks: Combining Statistical Approaches and Machine Learning Algorithms

Abstract

Net income is a key financial indicator that reflects the actual performance of the banking sector and its ability to achieve long-term profitability and sustainability. According to World Bank statistics, global banking profits exceeded $1.3 trillion, while the banking sector in the Middle East and North Africa (MENA) region recorded an annual net income growth of approximately 15% over the past decade. However, this sector faces fundamental challenges that threaten the stability of net income, most notably interest rate fluctuations, rising operating expenses, and changes in asset volume. In this context, this research aims to develop a predictive model for net income in banks by comparing the performance of two methodologically different models: the first is a traditional statistical model, represented by the Generalized Linear Model (GLM), and the second is a machine learning model, represented by the Decision Tree. The importance of this study lies in several aspects: first, it provides an intelligent analytical framework that enables the identification of the most influential factors shaping net income; second, it supports decision makers in banking institutions with accurate analytical tools to improve operational efficiency and rationalize expenses; Third, it bridges an existing research gap by integrating AI methodologies with classical statistical models in the context of financial analysis, rather than relying solely on traditional methods. The results are expected to significantly improve the accuracy of financial performance predictions, by 20% to 30% compared to traditional methods, making this hybrid approach an effective tool for formulating sustainable and resilient growth strategies in a volatile banking environment.