Leveraging Machine Learning for Enhanced Financial Inclusion: Revolutionizing Agricultural Banking in Central Africa

FOKAM NINYIM ASTRIDE MELVIN, VINCESS DONGMO

Keywords: Financial Inclusion, Agricultural Banking, Credit Scoring, Machine Learning, Explainable AI, Risk Assessment

TL;DR: credit scoring model using alternative data like mobile money transactions to enhance financial inclusion for rural farmers in Central Africa, comparing its performance with traditional methods.

Abstract:

Financial inclusion is a persistent challenge in Central Africa, especially for smallholder farmers who face limited access to credit and essential banking services. Agricultural banking in the region is hindered by unreliable data, inefficient risk assessment processes, and infrastructural gaps, restricting rural economic development and stability. This study investigates the transformative potential of machine learning (ML) technologies in agricultural banking focusing on improved credit scoring, robust risk management, fraud detection, and personalized financial products tailored to farmers’ unique needs.
Leveraging diverse datasets including transaction records, crop yields, weather data, and farmer demographics we develop and evaluate a suite of ML models, such as neural network,Decision Tree and ensemble methods (random forests, gradient boosting). These models are systematically compared to traditional benchmarks, including logistic regression and rule-based scorecards. Experimental results demonstrate that ML approaches significantly enhance loan approval rates, lower default risk, and boost operational efficiency when benchmarked against conventional methods.
Furthermore, the analysis highlights the critical role of explainable AI in fostering trust among stakeholders, ensuring regulatory compliance, and addressing ethical concerns such as data privacy and algorithmic fairness. The findings suggest that integrating machine learning into agricultural banking can drive inclusive growth in Central Africa by enabling scalable and tailored financial solutions. Future research should focus on overcoming scalability barriers and ensuring that the benefits of ML-driven banking are equitably distributed among all farming communities.

Submission Number: 12