Tablerank: a ranking algorithm for table search and retrieval

Abstract

Tables are ubiquitous in web pages and scientific documents. With the explosive development of the web, tables have be-come a valuable information repository. Therefore, effec-tively and efficiently searching tables becomes a challenge. Existing search engines do not provide satisfactory search re-sults largely because the current ranking schemes are inade-quate for table search and automatic table understanding and extraction are rather difficult in general. In this work, we de-sign and evaluate a novel table ranking algorithm – TableRank to improve the performance of our table search engine Table-Seer. Given a keyword based table query, TableRank facili-ties TableSeer to return the most relevant tables by tailoring the classic vector space model. TableRank adopts an innova-tive term weighting scheme by aggregating multiple weight-ing factors from three levels: term, table and document. The experimental results show that our table search engine out-performs existing search engines on table search. In addition, incorporating multiple weighting factors can significantly im-prove the ranking results.