Title: Study on news recommendation of social media platform based on improved collaborative filtering

Authors: Bin Wu

Addresses: The School of Network Communication, Zhejiang Yuexiu University, Shaoxing, 312000, China

Abstract: Aiming at the problems of low recommendation accuracy and low user interest in the existing methods, a news recommendation of social media platform based on improved collaborative filtering is designed. The initial key features of news data are determined, and the occurrence frequency of key features is counted by chi square, so as to realise feature extraction. First, we calculate the mutual information between different news data features, determine the correlation degree between features, and remove the data with similar features and low correlation degree. Then, the collaborative filtering algorithm is improved by adding timing update, trust and other data in collaborative filtering. Finally, the improved collaborative filtering algorithm is used to build a recommendation model, and the news data characteristics and user preference data are input into the model to complete the recommendation. The experimental results show that the news data recommended by the proposed method has high accuracy and high user interest.

Keywords: improving collaborative filtering; social media platform; news recommendation; active learning; covariance matrix; information gain.

DOI: 10.1504/IJWBC.2024.136675

International Journal of Web Based Communities, 2024 Vol.20 No.1/2, pp.27 - 37

Received: 10 Feb 2022
Accepted: 09 Jun 2022
Published online: 15 Feb 2024 *

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