Title: Top-k keyword search with recursive semantics in relational databases

Authors: Dingjia Liu; Guohua Liu; Wei Zhao; Yu Hou

Addresses: National Research Centre for Foreign Language Education, Beijing Foreign Studies University, Beijing, China; School of Information Science and Engineering, Yanshan University, Qinhuangdao, Hebei, China ' Department of Computer and Technology, Donghua University, Shanghai, China ' State Grid Heilongjiang Electric Power Company, Harbin, China ' School of Foreign Language Studies, Yanshan University, Qinhuangdao, Hebei, China

Abstract: Existing solutions for keyword search over relational databases focused on finding joined tuple structures from a data graph. We observe that such a graph using tuples as nodes and foreign-key references as edges cannot describe the joining connections between tuples within a single relation, and thus cannot support recursive query semantics over a relational database. To solve this problem, in our approach, we firstly model a weighted data graph considering both foreign key references and tuple joining connections within a single relation. Secondly, we discuss the ranking strategy for both nodes and edges supporting the recursive semantics by incorporating PageRank methods. Finally, an approximation algorithm as well as a top-k enumeration algorithm is presented by running Dijkstra algorithm based on dynamic programming strategy to enumerate result tuple trees. At the end of this paper, we conduct an experimental study and report the findings.

Keywords: relational database; keyword search; recursive semantics; graph; top-k; enumeration; shortest path; Steiner tree problem; group Steiner tree problem; PageRank; datalog.

DOI: 10.1504/IJCSE.2017.084699

International Journal of Computational Science and Engineering, 2017 Vol.14 No.4, pp.359 - 369

Received: 02 Oct 2015
Accepted: 02 May 2016

Published online: 21 Jun 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article