Effective Keyword Search in Graphs

TL;DR

This paper introduces two novel ranking methods for keyword search in graphs that incorporate node importance and bi-objective optimization, with efficient greedy algorithms validated on real data.

Contribution

It presents new ranking strategies for graph keyword search that consider node importance and optimize multiple criteria, addressing NP-hard problems with greedy algorithms.

Findings

Effective ranking results demonstrated on real datasets

Greedy algorithms achieve good efficiency and effectiveness

New methods outperform previous ranking approaches

Abstract

In a node-labeled graph, keyword search finds subtrees of the graph whose nodes contain all of the query keywords. This provides a way to query graph databases that neither requires mastery of a query language such as SPARQL, nor a deep knowledge of the database schema. Previous work ranks answer trees using combinations of structural and content-based metrics, such as path lengths between keywords or relevance of the labels in the answer tree to the query keywords. We propose two new ways to rank keyword search results over graphs. The first takes node importance into account while the second is a bi-objective optimization of edge weights and node importance. Since both of these problems are NP-hard, we propose greedy algorithms to solve them, and experimentally verify their effectiveness and efficiency on a real dataset.