Beyond Bag of Words: A New Model for XML Keyword Query

Abstract

Keyword search is an effective paradigm for information discovery and has been introduced recently to query XML documents. Effective keyword search of XML documents needs full understanding of the keyword query. Traditional bag of-words model cannot differentiate the roles of keywords as well as the relationship between keywords, thus is not proper for XML keyword queries. In this paper, we present a novel model specially designed for XML keyword query. The model takes a very different point of view on a keyword query: a keyword query is interpreted as a composition of several query units, each representing a query condition. We believe that this viewpoint captures the semantics of the query. To get an objective measure of the relevances of results with respect to the query, we devise a scoring method based on the proposed model that caters for query semantics as well as the structural properties of XML documents. Experimental results verify the effectiveness of our methods.