Sales forecasting of computer products based on variable selection scheme and support vector regression

摘要

Since computer products are highly replaceable and consumer demand often changes dramatically with the invention of new computer products, sales forecasting is therefore always crucial for computer product sales management. When constructing a sales forecasting model, discussing and understanding the important predictor variables can help focus on improving sales management efficacy. Aiming at to select appropriate predictor variable and construct effective forecasting model, this study combines variable selection method and support vector regression (SVR) to construct a hybrid sales forecasting model for computer products. In order to evaluate the feasibility and performance of the proposed approach, this study compiles the weekly sales data of five computer products including Notebook (NB), Liquid Crystal Display (LCD), Main Board (MB), Hard Disk (HD), and Display Card (DC) from a computer product retailer as the illustrative example. The experimental results indicate that the proposed hybrid sales forecasting scheme can not only provide a better forecasting result than the four competing models in terms of forecasting error, but also exhibit the capability of identifying important predictor variables. Furthermore, useful information can be provided by discussing the identified predictor variables for the five different computer products, thereby increasing sales management efficacy.