A Generic Ontology Framework for Indexing Keyword Search on Massive Graphs

Jiaxin Jiang, Byron Choi, Jianliang Xu, Sourav S. Bhowmick

Abstract: Due to the unstructuredness and the lack of schema information of knowledge graphs, social networks and RDF graphs, keyword search has been proposed for querying such graphs/networks. Recently, various keyword search semantics have been designed. In this paper, we propose a generic ontology-based indexing framework for keyword search, called Bisimulation of Generalized Graph Index ($\mathsf {BiG\hbox{-}index}$ ), to enhance the search performance. The novelties of $\mathsf {BiG\hbox{-}index}$ reside in using an ontology graph $G_{Ont}$ to summarize and index a data graph $G$ iteratively, to form a hierarchical index structure $\mathbb {G}$ . $\mathsf {BiG\hbox{-}index}$ is generic since it only requires keyword search algorithms to generate query answers from summary graphs having two simple properties. Regarding query evaluation, we transform a keyword search $q$ into $\mathbb {Q}$ according to $G_{Ont}$ in runtime. The transformed query is searched on the summary graphs in $\mathbb {G}$ . The efficiency is due to the small sizes of the summary graphs and the early pruning of semantically irrelevant subgraphs. To illustrate $\mathsf {BiG\hbox{-}index}$ 's applicability, we show popular indexing techniques for keyword search (e.g., $\mathsf {Blinks}$ and $\mathsf {r\hbox{-}clique}$ ) can be easily implemented on top of $\mathsf {BiG\hbox{-}index}$ . Our extensive experiments show that $\mathsf {BiG\hbox{-}index}$ reduced the runtimes of popular keyword search work $\mathsf {Blinks}$ by 50.5 percent and $\mathsf {r\hbox{-}clique}$ by 29.5 percent.