Comparison of Food Demand Forecasting Models Using Machine Learning Algorithms with Time Series Models

This paper focuses on the comparison of food demand forecasting models using machine learning algorithms and time series models. With the rapid development of technology and the increasing amount of available data for analysis, accurate demand forecasts are becoming a key factor for competitiveness in the food industry. The aim of this paper is to compare the performance of different food demand forecasting models based on machine learning algorithms and traditional time series models. To achieve this goal, the following objectives were set: to review the main time series models and machine learning algorithms, to identify their features and limitations in the context of food demand forecasting, to conduct a practical study on the data, to compare the effectiveness of the models on the given metrics and to formulate recommendations on the selection of the most appropriate forecasting method. Various time series models such as (S)ARIMA(X), Prophet, and machine learning methods including polynomial regression, SVR support vector method, random forest and gradient bousting were considered and analyzed in this paper. The features and limitations of each method in the context of food demand forecasting were identified.