A Service Recommendation Method Based On Trusted Similar Users

More and more users are involved in services computing. Services published by service providers are increasingly emerged. A service recommending system, which can exactly recommend a service satisfying the user’s preferred requirement from mass services, is required by people. However, most service recommending systems based on a single rate igore the user feedback information on each attribution of service, which makes recommendation not so exact. Also, fake evaluation may derease the trust degree of the recommended result.To solve the above problem, we propose a service recommendation method based on trusted similar users. The feedback information from users on services is first represented as multiple evaluation vectors including multiple attribution values. The user trust degree can be computed based on the herizonal and vertical processing of existed evaluation vectors. By deleting users with the low trust degree, the similarity between target user and other trusted users can be computed based on trusted evalution vectors. A neighbor user set can be determined based on the similarity and then the candidate recommendable services can be selected. Based on trusted evaluation vectors and target user preference weight, the candidate services are rated and the one with highest is recommend to the target user. A concrete case study is used to illustrate the feasible of our method. Moreover, the simulation is provided to illustrate the effect on untrusted users deletion and service recommendation by comparing result of deleting them or not.