IMPLEMENTASI METODE FUZZY TIME SERIES UNTUK PERAMALAN PENJUALAN PADA RESTORAN COCO BISTRO TANJUNG

Abstract

As competition in the culinary industry becomes increasingly intense, restaurant companies are required to focus more on product management and inventory control to maintain customer satisfaction and ensure smooth operations. Coco Bistro Tanjung, part of the Tropical Group, still relies on historical sales data without using scientific methods to forecast sales, which often leads to difficulties in determining the optimal amount of raw material inventory. This study aims to forecast sales for the upcoming period using the Fuzzy Time Series method, which is considered effective in addressing sales fluctuations. The research process begins with problem identification, a literature review, and the selection of the method, followed by the collection of sales data from Coco Bistro Tanjung. Forecasting is performed manually using the Fuzzy Time Series method, and the accuracy of the forecasts is evaluated using MAPE, MSE, and MAE. The results show that the Fuzzy Time Series method can provide accurate predictions for six main food categories, with the Balinese Main Course category achieving the best accuracy (MAPE 8.72%), classified as "Very Good," while the Pizza category showed the lowest accuracy with a MAPE of 38.75%, classified as "Fair." This research proves that the Fuzzy Time Series method is effective in forecasting sales trends for most food categories, which can help the company optimize inventory management and improve operational efficiency.