Benchmarking of Regression Algorithms and Time Series Analysis Techniques for Sales Forecasting

Abstract

Predicting the sales amount as close as to the actual sales amount can provide many benefits to companies. Since the fashion industry is not easily predictable, it is not straightforward to make an accurate prediction of sales. In this study, we applied not only regression methods in machine learning, but also time series analysis techniques to forecast the sales amount based on several features. We applied our models on Walmart sales data in Microsoft Azure Machine Learning Studio platform. The following regression techniques were applied: Linear Regression, Bayesian Regression, Neural Network Regression, Decision Forest Regression and Boosted Decision Tree Regression. In addition to these regression techniques, the following time series analysis methods were implemented: Seasonal ARIMA, Non-Seasonal ARIMA, Seasonal ETS, Non -Seasonal ETS, Naive Method, Average Method and Drift Method. It was shown that Boosted Decision Tree Regression provides the best performance on this sales data. This project is a part of the development of a new decision support system for the retail industry.