A Machine Learning Algorithm for Micro-credit Scoring

Abstract

The increasing adoption of mobile money services has paved the way for innovative financial solutions, particularly in the realm of micro-credit. This study introduces a Mobile Money Micro-Credit Scoring System designed to analyze transaction histories and assess creditworthiness for loan eligibility. The system addresses the evolving landscape of financial services, leveraging mobile money data, and data from a micro-finance institution to provide a comprehensive and data-driven approach to micro-credit evaluation in emerging economies, particularly Uganda. The borrowers targeted by micro-credits often lack traditional credit histories, necessitating a reliance on alternative data sources such as mobile phone usage and payment histories for utilities. To achieve this, we collaborate with a micro-finance institution, accessing diverse datasets to enhance the accuracy of credit assessments. This approach ensures a more robust and inclusive credit scoring model, mitigating the limitations faced by traditional credit assessment models. The study delves into the challenges faced by traditional credit assessment models, particularly in the context of emerging economies like Uganda, where access to formal financial systems is limited.By harnessing the vast transaction histories through mobile money platforms and additional data from micro-finance institutions, this system aims to bridge the gap in financial inclusion by offering timely and accurate credit assessments. The research incorporates a thorough literature review, exploring the existing landscape of mobile money, micro-credit, and credit scoring systems to establish the contextual framework for the proposed solution. Methodologically, the research employs machine learning algorithms and artificial neural networks to develop a robust credit scoring model. The focus on transaction history provides a dynamic and real-time assessment of an individual’s creditworthiness, contributing to a more inclusive and adaptive financial ecosystem. The aim of the Mobile Money Micro-Credit Scoring System is to achieve a reliable credit score that enables lenders to offer micro-credits to individuals who are traditionally excluded from formal financial systems. This approach balances risks and enhances accessibility to financial services. Notably, micro-loans are mobile money-based, inclusive of mobile apps for banks and micro-finance institutions, with a distinct emphasis on not being limited to specific mobile money systems like MTN or Airtel. This ensures a broader scope and applicability of the proposed solution