Short Papers
Deep Web, Web Form, Indexing, Query-by-similarity
Search engines do not provide specific searches for Web forms related to the Deep Web, in particular, similarity search. To deal with this lack, we propose a query-by-similarity system called WF-Sim, and this paper presents the indexing strategy adopted by WF-Sim for querying-by-similarity Web forms. It is centered on suitable index structures to the main kinds of queries posed on Web forms, as well as some optimizations in order to reduce the number of index entries. To evaluate the indexes performance, we ran experiments on two WF-Sim persistence strategies: file system and database. We also compare the performance of our indexes against the traditional keyword-based index, and the results were promising.
Willian Koerich, Ronaldo Mello