MODELLING OF BEER INDUSTRY DATA USING PRINCIPAL COMPONENTS AND TIME SERIES ARIMA METHODS

摘要

This study explores the interaction between the main market factors - digital and non-digital advertising, atmospheric temperature, price, distribution and others on the sales volume growth. Monthly observations for almost five years have been used for a big brewery brand in Bulgaria. Multivariate and stochastic ARIMA time series models with 82% and over 95% of coefficient of determination are constructed. The models are applied for a one-month ahead forecast of the beer sales volume.