Design and realization of personalized service in electronic commerce

Abstract

Classical collaborative filtering recommendation is the most successful recommendation algorithm in electronic commerce system application. However, along with the continuous increase of site structure, content complexity and user number, data is extremely sparse and the real-time property and recommendation accuracy of algorithm decrease significantly, even no any commodity can be recommended. This paper classifies the users in electronic commerce by collaborative clustering and carries out different page recommendations for different types of users to realize the personalized service in electronic commerce.