Towards an effective recommendation algorithm for e-commerce

• Main Author: Tang, Hao
• Other Authors: Zhang Jie
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/165948
• Institution: Nanyang Technological University

E-Commerce recommendation system has been extensively studied. A recommendation system makes it easier for customers to find products they like. Existing recommender system algorithm usually use collaborative filtering or content-based filtering. These methods still have some problems when executing it. Those problems may lead the lack of customer’s’ trust. In recent studies, there are still plenty of recommender system algorithms that not been used in shopping recommendation system. To improve the effectiveness of shopping recommendation system, we propose to use other algorithms to test if will overcome the current challenger faced by the collaborative filtering or content-based filtering. The proposed model contains two modules,(1) three baseline which is Neural collaborative filtering, Social collaborative filtering by trust and Diffnet++, that already existing in other fields, but not used in shopping recommendation system, followed by (2) to improve the accuracy of one of the baseline. We perform several experiments by using two datasets and the results show significant improvements as compared to previous works. Keywords: E-Commerce recommendation system, Neural collaborative filtering(NCF), Social collaborative filtering by trust and Diffnet++