Efficient fuzzy ranking for keyword search on graphs

초록

When compared with the traditional single-node results returned by search engines, keyword search over graphs is a new answering paradigm that brings new challenges to ranking. In this paper, we propose an efficient fuzzy-set theory based ranking measure called FRank. This measure captures the presence and relevance of query keywords and their query-dependent edge weights. It evaluates the query answer based on the distribution of keywords in the query and the structural connectivity between these keywords. Experimental results show that our proposed FRank measure led to superior performance when compared with traditional ranking measures. © 2012 Springer-Verlag.