Performance Analysis and Evaluation of Regression Models for Sales Forecasting

Abstract

Machine learning is transforming every sector of today’s world. By forecasting or forecasting sales, one can maximize the profits of business-to-consumer (B2C) models involving retail chains. This study examines aspects of predicting marketing trades by means of data. The results of this study will benefit store managers to provide a broad analysis of various forecasting methods for retail chain sales forecasting. This research study is to compare forecasting models that can analyze the seasonality of sales. Comparisons between different regression models such as linear regression, k-nearest neighbor, decision tree regression, random forest regression and XGB regression are performed and analyzed.