Collaboration reputation for trustworthy Web service selection in social networks

• We construct a Web service collaboration network.
• We propose a collaboration reputation concept.
• We present a trustworthy Web service selection method.

Traditional trustworthy service selection approaches focus the overall reputation maximization of all selected services in social networks. However, the selected services barely interact with each other in history, which leads to the trustworthiness among services being very low. Hence, to enhance the trustworthiness of Web service selection, a novel concept, collaboration reputation is proposed in this paper. The collaboration reputation is built on a Web service collaboration network consisting of two metrics. One metric, invoking reputation, can be calculated according to other service's recommendation. The other metric, invoked reputation, can be assessed by the interaction frequency among Web services. Finally, based on the collaboration reputation, we present a trustworthy Web service selection method to not only solve the simple Web service selection but also the complex selection. Experimental results show that compared with other methods, the efficiency of our method and the solution's trustworthiness are both greatly increased.