Research On WebService Recommendation Model Based On Social Network

In the recent years when distributed technology is rapidly developping WebService technology becomes more popular as a new type of distributed computing model,it has more use in Enterprise Management and Electronic Commerce.Then WebService Recommendation Model was born which has played a huge role in the promotion of WebService.The Tranditional WebService Recommendation System always builds on two kind of core algorithms:the Recommendation algorithm based on WebService function and the Recommendation algorithm based on Collaborative filtering.But there are still mounts of problems to solve in the two models.The performance of the Recommendation algorithm based on WebService function was always limited by a complete functional description ontology library as well as an accurate functional analysis matching algorithm,the Recommendation algorithm based on Collaborative filtering always has inaccurate result thanks to the poor scale of dataset.In order to avoid the huge impact of these issues on the recommendation system,this paper tries to use a new way,select the nowdays most popular technology-Social Network technology as a reference to complete the design of Recommendation algorithm.Among the complex relationships in social network,we choose the trusting relationship and the similarity relation to compute the influence of Social Network,proposed the WebService Recommendation Model based on Social Network.Firstly the research of the Measurement methods of the trusting relationship in social networks and the dissemination mechanism of trust in Social Networks is processed in this paper.In the direct trust relationship measurement,this paper choose the algorithm based on the set of customers' public friends.In the trust propagation algorithm design,this paper use the proportion based on the adjacent samples distance.Secondly in this paper the research of Measurement methods of similarity relations in social networks is processed and the disadvantage of the traditional similarity algorithm is analyzed,the characteristics of WebService and the traditional similarity algorithm are combined and the similarity calculation formula based on Qos attribute parameters is designed by calculating the similarity relationship between users and services.Then because the performance of the algorithm decreased when the dataset is becoming larger,this paper proposed a method based on the similarity of the sub groups,finally filter The similarity relationship between two nodes which has lower similarity result.At last two user nodes' computing result of similarity relationship are combined with the trust relationship and the Comprehensive recommendation based on social network relationship is computed.then calculate the final WebService's recommendation degree according to the attribute parameters of the the WebService quality evaluation model and the Comprehensive recommendation,give the result of recommendation modle by sorting the final WebService's recommendation degree.At the end of this paper, experiments for three parts of model are designed, compared and analyzed on the precision of recommendation result,the performance optimization in the recommendation accuracy. of WebService Recommendation Model designed in this paper is verified.