Demand forecasting: a comparison between the Holt-Winters, trend analysis and decomposition models

Abstract

Preliminary communicationIn food production industry, forecasting the timing of demands is crucial in planning production scheduling to satisfy customer needs on time. In the literature, several statistical models have been used in demand forecasting in Food and Beverage (F&B) industry and the choice of the most suitable forecasting model remains a central concern. In this context, this article aims to compare the performances between Trend Analysis, Decomposition and Holt-Winters (HW) models for the prediction of a time series formed by a group of jam and sherbet product demands. Data comprised the series of monthly sales from January 2013 to December 2014 obtained from a private company. As performance measures, metric analysis of the Mean Absolute Percentage Error (MAPE) is used. In this study, the HW and Decomposition models obtained better results regarding the performance metrics. Keywords: decomposition; demand forecasting; food industry; Holt-Winters; time seriesPredviđanje potražnje: usporedba između Holt-Winters modela, analize trenda i modela dekompozicije Prethodno priopćenje U industriji proizvodnje hrane, predviđanje vremena potražnje je bitno u planiranju proizvodnje kako bi se na vrijeme zadovoljile potrebe kupaca. U literaturi se koristi nekoliko statističkih modela za planiranje potražnje u industriji hrane i pića, a izbor najpogodnijeg modela od osnovnog je značaja. U tom kontekstu cilj je ovoga rada usporedba primjenljivosti modela analize trenda, dekompozicije i Holt-Winters (HW) modela za predviđanje vremenskih serija u potražnji đema i voćnih sokova. Obrađeni su podaci jednog privatnog poduzeća od siječnja 2013 do prosinca 2014. Uspješnost se odredila metričkom analizom Mean Absolute Percentage Error (MAPE) (srednji apsolutni postotak greške). U ovom su radu najbolji rezultati u planiranju potražnje postignuti modelima Holt-Winters i dekompozicije.

Developing a Multi-Stage Production Planning and Scheduling Model for a Small-Size Food and Beverage Company

“…The proposed MILP model is executed first for 12 months based on demands anticipated in our previous study [21] to test its effectiveness and after that, for the following two years, the MILP model is executed with actual demands.…”

“…This research aims to develop a production planning optimization model to improve the production of a real-world multi-stage and multi-line syrup and jam production company producing multiple products for both retail and wholesale with a limited capacity and whether the demand is either certain or uncertain for which forecasting is required. The model working on the predicted demand data considers tactical decisions, such as selecting the kind of product to be produced so that the demand can be satisfied [21]. To cope with such barriers, a specific mathematical model designed for production planning will be used and demonstrated based on the case company.…”

“…(i) The predicted demand amounts obtained using various forecasting methods such as the time series model considering seasonality and fuzzy applications were used in this model as well as the deterministic demand [21]. The amount of products to be made are planned on a monthly basis with a preference for early release prior to due date.…”

Time Series Long-Term Forecasting of per Capita Electricity Consumption for Bangladesh

“…In this study, actual historical demand data or consumption per capita of the past 43 years ranged from 1976/77 to 2018/19 (Bangladesh Power Development Board, 2019), has been used in different time series forecasting models (Tirkeş et al, 2017). The errors of the models are estimated to determine the best fit time series model (Towards Data Science, 2019) that coincides nearly with the actual historical data.…”

“…Previously, the time series model (Tirkeş et. al., 2017;Towards Data Science, 2019) has been applied for electricity demand or consumption forecasting for Bangladesh.…”