Research On Mashup Service Recommendation Method By Integrating User Interest And QoS

Mashup service is a new kind of Web service that mix different applications supporting Web API.Mashup service uses data which comes from external sources retrieved content, and combine more than one data source content to create entirely new and more value-added services. In recent years, with the continuous development of Mashup technology, the number and the typeof Mashup services are all in the high-speed growth. However, with the release of more and more Mashup services, it is easy for users to fall into an information overload predicament that similar to the Web service, users usually need to spend a lot of time to find expected Mashup services. Therefore, it becomes a challenging research problem how to recommend Mashup services, which meet user’s requirements and users are interested in according to the individual requierments of users, to improve user’s satisfaction.Currently, Mashup service recommendation methods are mainly based on content, collaborative filtering, quality of service, complex network and social network. There are some advantages and disadvantages for these methods.In this paper, aiming to Mash service recommendation, from the perspective of composite recommendation, we will integrate user interest and quality of service to solve mash service recommendation problem. The research results and the innovations of this paper are are as following two aspects:Firstly,to overcome the disadvantages of single recommendation method, this paper propose an effective Mashup service recommended method based on user interest and collaborative filtering.The method combines text-based similarity technology and social networking technology, computes the similarity between the services, establishes user interest model, and lastly gets user similarity model and service similarity model.The method will constructs user interest model and QoS prediction model of Mashup service, to compute user interest and QoS values of Mashup service, and then combines social network similarity of Mashup service, to generate Mashup service recommendation list and recommend Top-k Mashup service with high quilty of service to target user.Secondly, to solve the existing problem in Mashup services selection and recommendation based on quality of service, this paper proposes a Mashup service recommendation method based on information entropy multiple attribute decision making. The method will integrate user interest and quality of service, and use information entropy of multi-attribute decision making, to predict the probability distribution of user to target service and obtain the maximum probability score predictive value, thereby recommend Mashup service with high quality of service and interest to user.