DISCO: WEB SERVICE DISCOVERY CHATBOT

Abstract

An increasing number of companies offer their capabilities as Web services and publish them through public registries. Such Web services enable users to access and manipulate online data programmatically. Moreover, they can be deployed in service-based systems (SBS) or composite web applications and mashups to offer value-added services. A crucial step in deploying Web services is the Web service Discovery. Selecting the best Web service candidates to address the needs of service requesters (e.g., SBS designers) is an important task in SBSs and can influence the quality of the resulting composite web application. So far proposed solutions for addressing this challenge, which are mainly categorized into Syntactic-based and Semantic-based, are either too complex for end-users which hinders their large-scale adaptation or suffer from low precision and recall. Considering the recent trend of shifting the development activities more and more towards the end-users, we propose an approach based on chat-bot technology to allow end-users to select a set of best-fitting Web services according to their goals. Conversational Interfaces also known as chatbots aim at providing users a more natural framework to interact with web applications and devices. The advantage of DISCO compared to the state-of-the-art approaches is leveraging the natural language as the communication medium and alleviating the need for end-users to have technical knowledge about the service’s structure and location. We conducted sets of experiments by recruiting 8 test subjects to query the chatbot. The results achieve high precision and recall.