Recommending Web Services Using Crowdsourced Testing Data

With the rapid growth of Web Services in the past decade, the issue of QoS-aware Web service recommendation is becoming more and more critical. Web service QoS is highly relevant to the corresponding invocation context like invocation time and location. Therefore, it is of paramount importance to collect the QoS data with different invocation context. We have crawled over 30,000 Web servicesWeb services distributed across Internet. In this work, we propose to use crowdsourcing to collect the required QoS data. This is achieved through two approaches. On the one hand, we deploy a generic Web service invocation client to 343 Planet-Lab nodes and these nodes serve as simulated usersSimulated users distributing worldwide. The Web service invocation client is scheduled to invoke target Web servicesWeb services from time to time. On the other hand, we design and develop a mobile crowdsourced Web service tesing framework on AndroidAndroid platform, with which a user can easily invoke selected Web services. With the above two approaches, the observed service invocation data, e.g. response time, will be collected in this way. Then we design a TemporalQoS-aware Web service recommendation QoS-Aware Web Service RecommendationRecommendation Framework to predict missing QoS value under various temporal context. Further, we formalize this problem as a generalized tensor factorization model and propose a Non-negative Tensor FactorizationNon-negative tensor factorization (NTF) algorithm which is able to deal with the triadic relations of user-service-time model. Extensive experiments are conducted based on collected Crowdsourced testing data. The comprehensive experimental analysis shows that our approach achieves better prediction accuracy than other approaches.