Sales Forecasting Using GBDT Based Model And Data Mining Method

Abstract

Accurately predicting the sales of the mall can help companies adjust production strategies in a timely manner, improve production efficiency, and improve competitiveness. This article is based on the LightGBM model to realize Wal-Mart’s sales forecast. Due to the large amount of data in the data set given by the material and the relatively messy data types, we first perform feature processing on the original data, unify the abnormal data, and extract the data features, so as to obtain the processed data that can be used for modeling. In the use of grid search algorithm for parameter selection. Experiments show that the root mean square error of the LightGBM model is only 2.07, which has better predictive performance compared with the traditional linear regression model and SVM model.