Predictive Model for Financial Availability and Usability: Rural Population Segmentation

Abstract

Financial inclusion remains a critical challenge in global economic development, with approximately 1.7 billion adults’ worldwide still lacking access to basic financial services. Despite significant progress over the past decade, disparities in financial access persist across regions, income levels, and demographic groups. Recent years have seen significant advancements in “financial availability and usability” of banking and financial services prediction accuracy, thanks to AI-driven methods, especially those utilizing “Machine Learning” (ML) techniques. This study addresses the critical issue of financial inclusion by analyzing the impact of demographic variables (age, gender, education, income, occupation, and source of income) on financial availability and usability among rural populations & developing and validating predictive models for financial availability and usability, incorporating rural population segmentation. The most affected areas are categorized into two types through three different cluster analyses like K-means, “Hierarchical Clustering” (HC) and “Partitioning Around Medoids” (PAM) using all the 6 demographic variables. For highly affected areas, it is needed to predict “financial availability and usability”. The dat, aset of 102 data points pertaining to 70% of data set is utilized for training while that of 30% is utilized for testing. The collected data was analyzed using descriptive techniques and advanced statistical methods to identify trends and correlations.  Efficiency of the models is tested by cluster analysis, decision tree, regression model, and two “Artificial Neural Network” (ANN) models having hidden layer of one and two. All the models are using “Root Mean Squared Error (RMSE)” and “Mean Absolute Errors (MAE)” for performance evaluation. It is investigated that regression models show the lower RMSE and MAE results. Overall, this research highlights the potential advantages of implementing AI-driven methods in various regions and adds to the expanding body of work on using machine learning to accurately predict financial inclusion, a task that is both complex and vitally important.