Model Comparison for Predicting Beer Sales Volume: a Case Study of a Certain Brand of Beer

Abstract

The beer market is a sizable and important market in the alcoholic beverage industry. Accurate prediction of beer sales volume is crucial for entrepreneurs to employ in production planning and develop marketing strategies. As a result, this research aims to study models for predicting the sales volume of a certain brand of beer using multiple linear regression, decision trees, and random forest methods. The accuracy of the models is compared using the Root Mean Square Error (RMSE). The study utilizes monthly data from 2018 to 2021, consisting of 12 independent variables. The statistical analysis is conducted using the R programming language. The results reveal that the most suitable and efficient model for prediction is the multiple linear regression model, which uses a stepwise variable selection based on the Akaike Information Criterion (AIC). The significant independent variables influencing beer sales prediction include 1) the sales volume of the 1st competitor, 2) the third-tier selling price of the 1st competitor, 3) the third-tier selling price of the 2nd competitor, 4) the retail trade index, 5) the consumer price index for tobacco and alcoholic beverages, and 6) the population of Thailand aged 15 and above.