CLASSIFICATION OF PRODUCT PREDICATES BASED ON SALES RATE USING THE C4.5 DECISION TREE ALGORITHM IN RETAIL COMPANIES

General Background: Retail companies serve as crucial intermediaries between producers and end consumers, providing a wide range of products to meet daily needs. Specific Background: However, many retail companies encounter challenges in inventory management and stock provision, often stemming from insufficient analysis of sales and inventory data. Knowledge Gap: Existing research on inventory management in retail lacks a focus on predictive analytics techniques that leverage sales data to optimize ordering strategies. Aims: This study aims to identify sales patterns using the Decision Tree C4.5 algorithm, with the goal of predicting sales for various products to enhance ordering strategies. Results: Employing primary data collected via the company’s API and direct interviews with Order Management staff and the regional director of Jabodetabek, sales data spanning six months (November 2023 to April 2024) was analyzed using data mining techniques on the RapidMiner platform. The findings reveal that the Decision Tree algorithm effectively identifies product sales predicates, achieving a model accuracy of 96.20%. Novelty: This research introduces a data-driven approach to inventory management in retail, utilizing advanced decision tree algorithms for enhanced sales prediction. Implications: The implementation of the proposed model is expected to significantly improve the efficiency and effectiveness of the company’s ordering processes, ultimately leading to better inventory control and customer satisfaction.