Title: Market product demand forecasting method based on probability statistics and convolution neural network

Authors: Yingji Cui

Addresses: School of Mathematics and Information Science, Xianyang Normal University, Xianyang 712000, China

Abstract: Due to the high proportion of redundant data and high prediction error rate in traditional market product demand prediction methods, this paper proposes a market product demand prediction method based on probabilistic statistics and convolutional neural network. Design the market product demand data acquisition module, collect the market product demand data, and use the improved near record sorting algorithm to clean the collected data. The statistical model of market product demand change probability is constructed to obtain the demand change probability, and the data cleaning result and demand change probability are taken as the input of the convolutional neural network model, so as to obtain the market product demand forecast result. The experimental results show that the maximum redundant data of this method accounts for only 5.9%, the prediction error rate varies between 1% and 3%, and the average prediction time is only 0.22 s. The practical application effect is good.

Keywords: probability statistics; convolution neural network; market products; demand forecasting; sorted neighbourhood algorithm.

DOI: 10.1504/IJPD.2021.116154

International Journal of Product Development, 2021 Vol.25 No.2, pp.187 - 199

Received: 26 Nov 2020
Accepted: 14 Mar 2021
Published online: 12 Jul 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article