Product Recommendation in e-Commerce Using Direct and Indirect Confidence for Historical User Sessions

Product recommendation is an important part of current electronic commerce. Useful, direct and indirect relationships between pages, especially product home pages in an e-commerce site, can be extracted from web usage i.e. from historical user sessions. The proposed method introduces indirect association rules complementing typical, direct rules, which, in the web environment, usually only confirm existing hyperlinks. The direct confidence, the basic measure of direct association rules, reflects pages’ co-occurrence in common user sessions, while the indirect confidence exploits an additional, transitive page and relationships existing between, not within sessions. The complex confidence, combining both direct and indirect relationships, is engaged in the personalized process of product recommendation in ecommerce. Carried out experiments have confirmed that indirect association rules can deliver the useful knowledge for recommender systems.

• Type: Conference
• Year: 2004
• Where: DIS
• Authors: Przemyslaw Kazienko
• Keywords: Direct, DIS 2004, Indirect Association Rules, Indirect Relationships