Walmart Sale Forecasting Model Based On LightGBM

Accurate sales forecasts can help businesses make reasonable sales plans. In this paper, I am going to build a model for sales forecasting. First, I need to preprocess the historical data, find the obvious feature structure in the data, and eliminate the unreasonable data. I propose a GBDT-based LightGBM model, the Root Mean Square Error of this model is 2.09, which shows better predictive ability compared with linear regression model and SVM model.