SWEE: Approximately searching web service with keywords effectively and efficiently

With the wide use of web services over the world, an effective and efficient search engine becomes necessary. In this paper, we analyzed the limitation of search method using traditional UDDI, and make it in IR-style and take advantage of approximate string matching. Motivated by the feature of WSDL files which define the interfaces of web services, we make operations defined in WSDL files which compose web services as the base units for searching and organize all information of operations and corresponding components as documents, which will facilitate IR-Style keyword searching. In order to improve the precision of searching, we enable users to specify weights in their queries. By defining our ranking function, we sort all the results and present top n ones to users. We also build inverted indexes on operations and corresponding documents to improve query efficiency. In addition, we carry out simple semantic analysis on WSDL files to acquire better accuracy. At last, not only have we tested the efficiency of query after introducing approximate string matching, but also have we verified the effectiveness of the query results.