Title: Intelligent recommendation of e-commerce products combining topic features and attention mechanisms

Authors: Xingchang Yang

Addresses: School of Information Engineering (School of Software), Henan University of Animal Husbandry and Economy, Zhengzhou, 450044, China

Abstract: Today's society development relies on big data. Users evaluate via text review data, and platforms obtain feedback from it. To enhance e-commerce platform recommendation services, this study presents an intelligent e-commerce product recommendation system integrating topic features and attention mechanisms. It uses the keyword-weighted Shark-PageRank algorithm for topic web crawling, a template-based method for feature extraction, and a hybrid attention mechanism [bi-long short-term memory+, (Bi-LSTM+)] model to calculate sentiment. Experiments show the keyword-weighted algorithm's accuracy reaches 84.5%, and the extraction method's, 92%. The system's real-time score averages 9.9, lag percentage is below 1%, and security is 93.3%, meeting standards. It has guiding significance for e-commerce platforms.

Keywords: feature extraction; attention mechanism; e-commerce products; sentiment analysis; intelligent recommendation.

DOI: 10.1504/IJWET.2025.151162

International Journal of Web Engineering and Technology, 2025 Vol.20 No.4, pp.422 - 440

Received: 24 Jul 2024
Accepted: 17 Mar 2025
Published online: 15 Jan 2026 *

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