Reputation model for C2C commerce with filtering merchandise categories

With the large number of buyer and seller agents in C2C commerce, the process managing to establishing trust relations using previous repeated direct interactions is difficult to achieve. In a loosely connected e-market, the best way to establish trust relations is probably with the help of the commendation coming from the buyer agents who have traded with the same seller agent. Furthermore, the effectiveness of the commendation is partly determined by the similarity of contents between historical and present trades — context relevance which is identified through the merchandise categories in this paper. With the concepts — buyer's merchandise-categories reputation and buyer's recommended reputation filtering merchandise categories — being proposed, and the experience cumulative function used for calculating the buyer's recommended reputation and the equation used for aggregating seller's reputation being designed, a reputation system based on SPORAS is improved. Finally, compared with SPORAS, the benefits of filtering merchandise categories to manage reputation in C2C commerce are analyzed according to the excellent experimental results shown.