Assessing Digital Financial Inclusion of Women Entrepreneurs Using Artificial Intelligence Techniques

Abstract

Digital financial inclusion (DFI) of women entrepreneurs remains a critical yet underexplored challenge in developing economies. This paper presents a comprehensive AI-driven framework to assess, classify, and predict the digital financial inclusion status of women entrepreneurs across urban, peri-urban, and rural segments. Leveraging a curated dataset of 4,820 respondents drawn from six Indian states, we apply five machine learning techniques- Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting (GB), Artificial Neural Network (ANN), and a hybrid Stacked Ensemble (SE)- to evaluate financial access, digital literacy, product adoption, and socioeconomic barriers. The proposed stacked ensemble model achieves an accuracy of 94.7%, F1-score of 0.943, and AUC-ROC of 0.971, outperforming all baseline models. Key features identified include mobile internet access (IG=0.412), prior banking history (IG=0.387), educational attainment (IG=0.361), and awareness of digital credit products (IG=0.318). Findings reveal a 34.2 percentage-point rural–urban DFI gap and demonstrate the practical utility of AI for policy targeting. This study contributes a validated index (WDFI-Index), an open benchmark dataset, and actionable recommendations for policymakers, financial institutions, and development organizations.