Sales Forecasting Analysis Using Fuzzy Time Series and Simple Linear Regression Methods at Toko Ari

Introduction: Forecasting, often referred to as prediction, can actually help assess conditions or predict future sales. In the business world forecasting is crucial because it can help companies plan their future operations especially when faced with sudden increases and decreases in sales and stockpiles. Especially in retail forecasting is extremely helpful in purchasing merchandise, managing inventory in the warehouse, and reducing losses due to changing customer preferences. Ari's shop, located on Jalan Raya Samu, Singapadu Kaler, Gianyar, Bali, also experiences increases and decreases in monthly sales. Therefore, it is hoped that this sales forecasting can help maintain more stable and smooth operations. Methods: This study used two methods to forecast sales: Fuzzy Time Series (FTS) and Simple Linear Regression (SLR), to predict figures from Ari's shop's monthly sales data. Both methods use the same dataset, which is Ari's Store sales data for 13 months, from January 2024 to January 2025. The forecast results are then compared using the Mean Absolute Percentage Error (MAPE), which measures the model's accuracy in predicting results. Results: Based on the sales forecasts performed, both models produced fairly accurate predictions due to their low MAPE values, below 10%. Of the two methods, Simple Linear Regression provided more accurate results with a MAPE of 3.57%. Meanwhile, the Fuzzy Time Series method produced a MAPE of 5.53%. This difference in values indicates that the linear regression model is more appropriate for Ari's Store sales data, especially since the data pattern tends to follow a linear trend.