Research on the Application of LMM Model with Fusion Feature Extraction in Forecasting the Marketing Quantity of E-commerce Enterprises
Accurate sales forecasts can help e-commerce enterprises develop appropriate marketing plans to reduce bus-iness losses.However,traditional prediction models are no longer sufficient to meet the needs of e-commerce enterpri-ses in the era of rapid development of big data.This study is based on feature extraction and Fourier transform processing of sales data to construct a Linear Mixed Model for predicting the marketing quantity of e-commerce enterprises.Con-duct experiments to analyze the utility of the prediction model.The experiments show that the MAE of the neural network and support vector machine models is 0.04927154 and 0.04847517,while the MAE of the linear hybrid model based on feature extraction is smaller,0.04032,indicating better model quality and performance.When it comes to the evaluation indicators of RMSE,MAE,RMSE,and RMSPE,the errors of the three models are all within 0.1.E-commerce mar-keting quantity data has time series characteristics,but linear mixed models are more effective in solving time series problems,and therefore can process e-commerce data more effectively.
NETL
NSTL
万方数据
