APPLICATION OF FUZZY LOGIC IN PREDICTING GROCERY STORE REVENUE

Abstrak

Objective: This study aims to predict grocery store income accurately using the Mamdani fuzzy logic method to address challenges caused by price fluctuations and market uncertainty in retail operations. Method: Data were collected through direct observation at a grocery store in Sidoarjo over one month, including variables such as the number of items sold, total price, and operational costs. Each variable was classified into fuzzy sets using triangular and shoulder-shaped membership functions. The fuzzy inference system consisted of 27 if–then rules, with output values determined through the Weighted Average defuzzification method. Results: Based on input data of 1,004 units sold, a total price of Rp. 14,124,500, and operational costs of Rp. 4,500,000, the system successfully predicted an income of Rp. 10,000,000. Evaluation using Mean Absolute Percentage Error (MAPE) indicated an error rate of 0%, signifying exceptional predictive accuracy. Novelty: The study demonstrates that fuzzy logic can serve as a highly effective decision-support tool for income prediction in small-scale retail businesses, providing a reliable framework for managing financial uncertainty.

Kata Kunci