QoS-Aware Web Service Recommendation using a New Collaborative Filtering Approach

Abstract

Predicting Quality of Service (QoS) of web services through recommendation systems for accessing the information required by users is possible. But the utmost challenging in such systems is how to find the optimal service corresponding with the needs and requests of the users among a variety of web services. Satisfactory in quality of web service depends on its performance and this performance is measured through QoS. Collaborative Filtering (CF) plays important role in recommendation system. In general performance of CF is manipulating of a bunch of candidates with similar functionalities in predicting data distinguished by them, which is calculated by Pearson Correlation Coefficient (PCC). In this paper, we introduce a new collaborative filtering approach for predicting of the values of QoS of web services, also we introduce a Web service recommendation by taking in account of advantages of using of candidates foregoing experiences. In purpose of increasing accuracy of CF, values of similarity of main item and average of items were added to CF of candidates. This method which is called New Pearson Correlation Coefficient (NPCC), which is a combination of user-based and item-based methods. In purpose of investigating the accuracy of our proposed predicting of QoS, we have used a subset of the WSdream dataset to predict the QoS values. The outcomes of using our proposed method indicate the better performance and results compared to other methods outcome.