Machine Learning Modeling for Flexible Management in Financial Inclusion Assessment

Abstract Microfinancial inclusion (MFI) is key to ensuring fair access to financial services in India, especially for marginalized groups. However, existing inclusion metrics often fail to capture regional disparities and lack adaptability to evolving socioeconomic contexts. This study proposes a new MFI index based on machine learning to support flexible management through a data-driven, adaptable policy framework that responds to regional differences and changing socioeconomic conditions. Using financial and demographic data from 2016 and 2022, the index construction follows a five-step framework: parameter selection via XGBoost, sub-indexing, rank order centroid (ROC) weighting, aggregation using weighted quadratic mean (WQM), and classification. The index shows significant differences in inclusion trends across Indian states. States like Jharkhand and Manipur exhibit high-growth rates, but their overall inclusion levels remain low. Meanwhile, states such as Punjab and Uttarakhand faced stagnation or decline. This highlights issues such as policy fatigue and limited infrastructure. The southern region shows higher MFI dominance than the northeastern and eastern regions. The results underscore the importance of targeted strategies that focus on specific regions, rather than implementing one-size-fits-all national policies. By combining data analysis with adaptable governance methods, this study provides a scalable approach to continuously monitor and enhance financial inclusion. This approach provides policymakers with useful tools to identify structural weaknesses and distribute resources more effectively.