Research On Web Service QoS Prediction And Active Recommendation Algorithm Based On Collaborative Filtering Specialitycomputer Technology

With the increasing number of Web services in the network,QoS(Quality of Service)-driven Web service recommendation has become a hot topic in the field of service computing.The active recommendation technology of Web service can free user from the massive service,solve the information overload problem of Web service,and bring convenience for user to provide more suitable and high-quality Web service.The main process is to predict the missing QoS value of the active user by mining the user's history QoS records,and select the web services with better QoS to recommend to active users according to the predicted QoS values.This paper focuses on the following problems in QoS prediction using traditional collaborative filtering algorithm.Problem 1: When faced with sparse matrix,the prediction accuracy QoS Web service is not ideal.Problem 2: When there are unreliable data provided by untrustworthy users in the historical QoS data,the accuracy of QoS prediction will be greatly affected.The main contribution of this study includes:1.Aiming at the problem of inaccurate QoS Prediction results for Web service in sparse matrix,a k-means clustering method is presented to predict Web service QoS through Principal Component Analysis(PCA).This algorithm enriches the QoS information and improves the prediction accuracy by PCA filling the original sparse user item matrix.2.Aiming at the influence of unreliable data provided by untrustworthy users on QoS prediction in historical QoS data,this paper presents a K-Means clustering prediction method considering user geographic location information and credibility.This method synthetically considers the user's geographic location information and credibility,reduces the influence of unreliable data and user context(such as user location,user network status,etc.)on QoS prediction,and improves the accuracy of prediction.3.In order to improve the utilization of valuable information and the adaptability of QoS prediction algorithms in different datasets,this paper presents a QoS prediction method based on mixed user information and Web service item information.Considering the location information and credibility information of users and Web services,this method improves the utilization of valuable information in the absence of similar users or similar service items.Meanwhile,we adjust the dependence of prediction algorithms based on user and Web service terms on different datasets,avoid the influence of different data correlation characteristics on QoS prediction accuracy and improve the prediction accuracy.Comprehensive experiments are conducted on WS-Dream dataset from the real world.Experimental results show that compared with the traditional QoS prediction method,the prediction results of the proposed algorithm are more accurate and timely,and the QoS prediction accuracy of the algorithm is up to 14.47%.