Predicting Determinants of Financial Inclusion through Socioeconomic Indicators of Egypt: A Comparative Study of Traditional and Artificial Intelligence (AI)-Based Approaches | ||||
| مجلة البحوث المالية والتجارية | ||||
| Volume 26, Issue 3 - Serial Number 1, July 2025, Page 1081-1124 PDF (659.99 K) | ||||
| Document Type: المقالة الأصلية | ||||
| DOI: 10.21608/jsst.2025.394894.2049 | ||||
 View on SCiNiTO | ||||
| Author | ||||
وائل مصطفى حسن محمد 
| ||||
| كليه العلوم الماليه والإداريه -جامعه فاروس بالإسكندريه | ||||
| Abstract | ||||
| This research explores the multidimensional determinants of financial inclusion in Egypt through the application of traditional econometric techniques namely, Bootstrap Regression Analysis as well as state-of-the-art artificial intelligence (AI) algorithms that include Random Forest (RF) and Support Vector Regression (SVR),using published data by central bank of Egypt and World Bank from 2016 till 2023, this study puts in perspective the fast transition in terms of certain socioeconomic drivers: Human Development Index (HDI), Growth of GDP, unemployment rate, and concentration of economic sectors represented by (HHI). According to the bootstrap regression effects, HDI has the strongest positive effect on financial inclusion, emphasizing the importance of the development of human capital in facilitating the access to financial services. On the contrary, high HHI, which means low market competition, indeed poses significantly negative effects on financial inclusion, suggesting the significance of financial sector openness and competition. AI models outperformed the traditional approach in predictive performance with SVR achieving the greatest (R² ≈ 0.99, MAE ≈ 0.15) and being closely followed by RF. Meanwhile, for inference, although bootstrap regression seemed to have some kind of forecast superiority, it has strong inferential power, which was able to contribute to counter some sort of classical assumption violations via resampling. Permutation-based feature importance indicated the differentiation of models’ insights: unemployment was identified by SVR as highly significant, whereas RF confirmed HDI and GDP as relevant. These differences suggest that alternative methodological paradigms may provide alternative insights into the structural determinants of financial inclusion | ||||
| Keywords | ||||
| Financial Inclusion; Bootstrap regression; AI; Machine Learning; Random Forest (RF) and Support Vector Regression (SVR) | ||||
|
Statistics Article View: 315 PDF Download: 222 |
||||
