A data-centric and machine based approach towards fixing the cold start problem in web service recommendation

Web services are independent, modular and autonomous pieces of software offering functionality that help an organization achieve reduced development cost. However, the issue of selecting a web service from a set of similar services is non-trivial. Recently, service recommendation via collaborative filtering has started to receive attention in the research community as a criterion for selection. However, a problem with such a method is `cold start´, wherein some of the users must invoke certain web services, so that other services can be recommended. In this paper, we propose a data driven approach towards web service recommendation to solve the cold start problem. To demonstrate the viability of the proposed technique in real time applications, we experiment with a real world dataset consisting of 150 web services invoked by 100 users around the globe. Moreover, as a proof of concept, we have developed a web based prototype capable of predicting QoS values and recommending services in real time. We present and discuss our findings in the result section.