Theses and Dissertations - UTB/UTPA
Date of Award
Master of Science (MS)
Dr. Artem Chebotko
Dr. Xiang Lu
Dr. Bin Fu
We formulate and tackle a flexible and useful query, namely k-nearest keyword (k-NK) query, which can identify the relationship between vertices (or keywords) in an RDF graph, where users are only required to specify two query keywords q and w (without knowing the domain knowledge). In particular, a k-NK query returns k closest pairs of vertices (ui; vi) in the RDF graph such that vertices ui and vi contain keywords q and w, respectively, and vi has the smallest (shortest path) distance to ui (i.e., vi is the nearest neighbor of ui). In order to efficiently answer k-NK queries, in this paper, we propose three efficient query answering techniques that utilize effective pruning strategies and cost-model-based indexing mechanisms. We also confirm the effects of our proposed approaches on real and synthetic RDF data sets through extensive experiments.
University of Texas-Pan American
Copyright 2012 Eugenio De Hoyos. All Rights Reserved.