Title: Hotspot extraction method of multimedia network public opinion based on neural network

Authors: Mi Tian

Addresses: School of Design and Art, Changsha University of Science and Technology, Changsha 410114, China

Abstract: In order to overcome the problems of traditional methods, such as poor accuracy of hotspot extraction and poor information recall rate, this paper proposes a method of hotspot extraction of multimedia network public opinion based on neural network. The hot information of multimedia network public opinion in Weibo database is collected by association rules, and the redundant information of multimedia network public opinion hotspot is denoised. Relief algorithm is used to realise the relevant feature screening of public opinion hot information, and the neural network structure is used to calculate the weight threshold of public opinion words to construct the feature target extraction function of public opinion hotspots, so as to realise the multimedia network public opinion hotspots extraction. The results show that the method presented in this paper can improve the accuracy of hotspot information extraction and reduce the CPU consumption during extraction.

Keywords: multimedia network public opinion; target extraction function; TF-IDF algorithm; neural network training.

DOI: 10.1504/IJWBC.2022.125491

International Journal of Web Based Communities, 2022 Vol.18 No.3/4, pp.301 - 317

Received: 29 Jun 2021
Accepted: 08 Nov 2021
Published online: 12 Sep 2022 *

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