Quality-driven Extraction, Fusion and Matchmaking of Semantic Web API Descriptions

The composition of Web APIs provides a great opportunity to Web engineers that can reuse existing software components available on the Web. Finding the best API, ful lling a set of user requirements, among the many described on the Web is a key step in order to de- velop an e ective Web application; however, Web engineers have little support in solving this problem due to poor search mechanisms and to the heterogeneity of sources and descriptions. Semantic technologies and matching algorithms provide accurate methods to match user require- ments against a set of descriptions. Nonetheless, semantic descriptions of APIs are not available in practice. In this paper, we propose a method to extract information on Web APIs published in several Web sources and create semantic descriptions that can be then fused to deliver com- prehensive descriptions associated with APIs. During the extraction pro- cess, we take into account that collected information has di erent levels of accuracy, currency, and trustworthiness to state a con dence level of the results. The method is based on the evaluation of the quality of the involved sources, the extracted values, and the overall descriptions. The resulting semantic descriptions are then matched with expressive user requirements to address the API selection problem.