MACHINE LEARNING RISK PREDICTION FOR SUSTAINABLE MSMES ACROSS ECONOMIC, SOCIAL, AND ENVIRONMENTAL DIMENSIONS

Abstract

This article develops a machine learning-based framework for predicting economic, social, and environmental risks among sustainable micro, small, and medium enterprises. The study uses an integrative review and scenario-based modeling approach to translate sustainability and SME digital transformation literature into a practical risk architecture. The proposed model combines financial vulnerability indicators, workforce stability, supply-chain exposure, energy intensity, waste practices, and digital maturity. The results show that machine learning can support early warning and targeted assistance when the model is explainable, data requirements are proportional, and outputs are linked to advisory interventions rather than punitive scoring. The article argues that sustainable MSME analytics should prioritize interpretability, low-cost data collection, and ecosystem support. The proposed framework can guide policy makers, lenders, cooperatives, and business-development agencies in designing risk dashboards that protect resilience while avoiding exclusionary algorithmic decisions.