Forecasting the amount of food to order ingredients for a restaurant by Time Series Method

Abstract

This research presents a forecast of the number of orders for food items with the most orders from customers of a case study restaurant. To be used to order raw materials, the data used in the forecast was collected from June 2023 to December 2023, a total of 6 months ( 29 weeks) , which includes trend components. It presents four forecasting techniques as following: simple moving average method (3, 6 and 9 periods) , simple exponential smoothing method, Holt's double exponential smoothing method and Simple linear regression model method. For choosing a forecasting technique to predict the number of raw materials, the error values from the forecast are used to decide on the forecasting method using the minimum of Mean Absolute Percentage Error ( MAPE) value. The results of the research found that the 9- period moving average method give the minimum error. Therefore, the obtained forecast values were used to order raw materials in the next period.