A novel service recommendation approach considering the user’s trust network

Abstract

Web services are ever increasingly published on the network as core components of Service-oriented architecture (SOA). An attendant problem is how to help users select their satisfied services that meet their functional and non-functional requirements from the mass services. Service recommendation technology is adopted and studied as an effective approach currently. This paper focuses on the user’s trust network, where the users share their experience and rating for the invoked services. To attack the data sparsity and cold-start problems in the user-service rating matrix, an improved random walk algorithm is proposed. Firstly, we employ the non-negative matrix factorization method to compute the similarities between users and services separately. Then our method introduces the trust relationship in iterations of the random walk to select the trust users accurately. At last, the real dataset is used to validate our approach. Experimental results show the effectiveness of our approach compared with the state-of-art algorithms.