A New Sales Forecasting Method for Industrial Supply Chain

Abstract

With the continuous innovation of computer technology, it can solve the problems of low accuracy, non-intelligence, and inability to process complex samples in the sales forecasting methods of industrial supply chains. This paper proposes a sales forecasting method for the industrial supply chain based on the Gaussian mixture model. By analyzing the characteristics of the original sales data of the industrial supply chain, the eigenvalue correlation ranking vector is generated. Then predict the parameters such as the number of clusters in the Gaussian mixture model. By comparing the accuracy of the prediction results, the recall rate and the Fvalue, the eigenvalues, and the number of clusters that can achieve better prediction results are determined. This paper compares the Gaussian mixture model with the artificial neural network model and the convolutional neural network model on the original sales data set of the same industrial supply chain. The experimental results show that, compared with the artificial neural network model and the convolutional neural network model, the method has better performance in all three indicators, and can better predict sales transactions.