A BOT-ENABLED APPROACH FOR WEB SERVICE DISCOVERY

Abstract

With the wide adoption of service-oriented architectures, many service-based systems (SBSs) and mashups composed of Web services have been developed. Meanwhile, web service discovery, as a crucial component to guarantee the quality of Web service composition in SBS, gained increasing attention. Web service discovery aims to select the set of best service candidates according to the service requester’s requirements. Existing syntactic and semantic-based approaches are proven to have low recall and precision or expecting high level of expertise from the service requesters. In this paper, we propose an approach for Web service discovery based on chatbots which obtains the best suited set of services to address the goals of service requesters. Chatbot-based interfaces are the emerging trend to provide a more natural human-device interaction. Compared to the existing web service discovery techniques, our approach alleviates the need for service requesters to have prior knowledge about the service location or its structure and enables them to identify services by leveraging natural language. We evaluated the approach through the collection of 56 queries from the test subject group on our repository of existing web service descriptions and report on quality metrics. The results shows high precision and recall.