Research on Web-of-Trust-Based Personalized Seller Recommendation Algorithm for E-Commerce

Abstract

Trust evaluation and personalized recommendation are two challenge problems in large scale e-commerce applications. In this paper, we present a novel web-of-trust-based personalized seller recommendation algorithm. Starting from transaction-related data and trading relationship, the buyer-buyer web-of-trust and buyer-seller web-of-trust for recommendation and related concepts are formally defined. The method of constructing the web-of-trust is discussed. The algorithms computing direct trust and recommending trust between buyers based on web-of-trust are presented, and the algorithm for computing direct trust of the buyer in the seller is described. Aiming at the needs of the user intending to buy a specific product, the web-of-trust-based personalized seller recommendation algorithm is proposed, which can be used to find the most trustworthy seller selling the specified product for the buyer. The algorithms for computing direct trust of the recommender in target seller and comprehensive trust of the requester in target seller are designed. Using an example, the process of generating seller recommendation is described, and the recommendation results are analyzed.