Research On Trustworthy Service Recommendation In Big Data Environments

Service recommendation is an important part of service selection. With the development of social networks are increasing so quickly that service selection based on user demand are facing challenges in the big data environments. One effective way to solve the above problem is to divide the social network by the community detection method. But existing method has a high computation cost. Meanwhile, it doesn’t consider that user trust may affect the effect of community detection method and malicious users are growing in a high speed which will also reduce the accuracy of the recommendation result. How to dynamically add new users and remove the malicious users are also difficult problems.In order to advance effective and trustworthy community in big data environments, a service recommendation method based on trustworthy community in big data environments is proposed with integration of social network theories and trust theories. This paper’s main work is as follows:(1) Trust relationship is established based on available trust models and trust of users on service providers is calculated. The time-varying and dynamic characteristics of trust are considered. This paper also proposes an optimization method of trustworthy community, including the addition of new users and the deletion of malicious users. The method of adding new users is based on user’s interests and preferences. The method of deleting malicious users is based on trust relationship and the similarity of users and neighbors. The simulations reveal that the proposed method can avoid malicious attack from service recommenders effectively.(2) With construction of trust relationship, trustworthy community is built. On the basis of the construction of trustworthy community, this paper proposes a trustworthy community detection algorithm. The algorithm is used to divide massive users effectively. A service recommendation method based on the trustworthy community is proposed with the employment of MapReduce. The paralleled collaborative filtering recommendation method is also proposed to improve the efficiency of the algorithm. The simulations reveal that the proposed method can improve the efficiency of service recommendation compared with other traditional recommendation methods.(3) This paper designs a service recommendation system based on the theory and method above, and shows the design of the system including demand analysis, general design, module design in detail and implement process. The test results show that the system has high stability and efficiency.