A FUZZY LOGIC-BASED MATHEMATICAL MODEL FOR AGRI-E-COMMERCE FINANCING READINESS AND SUPPLY-CHAIN RESILIENCE: R-BASED SIMULATION WITH MATLAB VISUALIZATION

Source abstract

Small agricultural enterprises use e-commerce to reach buyers, document sales, and shorten market channels, yet many remain excluded from formal finance because lenders cannot translate fragmented digital and supply-chain information into credit decisions. This study develops a Mamdani fuzzy inference model that integrates digital transaction strength, financial reliability, and supply-chain resilience into an Agri-E-Commerce Financing Readiness Index. Each input is normalized on a 0-100 scale and represented by low, medium, and high linguistic membership functions. A 27-rule base converts overlapping memberships into low, moderate, or high readiness, followed by centroid defuzzification. The model was specified in base R for use in RStudio and evaluated through a reproducible Monte Carlo sample of 150 agri-enterprises. Simulated observations reflected heterogeneous digital adoption, repayment capacity, and logistics performance. The model produced a mean predicted readiness score of 51.08 compared with an observed benchmark mean of 50.73. Predictive validation yielded RMSE = 5.623, MAE = 4.511, and R² = 0.919. Rule inspection showed that high digital sales alone could not offset weak financial reliability or a fragile supply chain; readiness increased most consistently when at least two dimensions were medium-to-high and none was critically low. The model therefore provides an interpretable screening layer for agricultural lenders, e-commerce platforms, cooperatives, and policymakers. It is not a substitute for regulated underwriting, but it demonstrates how imprecise operational evidence can be converted into an auditable readiness score. Future research should estimate membership thresholds from field data and test fairness across farm sizes, regions, and commodity groups.