Recommendation techniques aim to support the users in their decisionmaking while the users interact with large information spaces. Recommendation has been a hot research topic with the rapid growth of information. In the field of services computing and cloud computing, efficient and effective recommendation techniques are critical in helping designers and developers to analyse the available information intelligently for better application design and development. To recommend Web services that best fit a user’s need, QoS values which characterize the nonfunctional properties of those candidate services are in demand. But in reality, the QoS information of Web service is not easy to obtain, because only limited historical invocation records exist. So in this project present a model named CLUS for reliability prediction of atomic Web services, which estimates the reliability for an on going service invocation based on the data merged from previous invocations. Then aggregates the past invocation data using hierarchy clustering algorithm to achieve better scalability comparing with other current approaches. In addition, the paper proposes a modelbased collaborative filtering and location based recommendation approach based on supervised learning technique and linear regression to estimate the missing reliability values.
展开▼
