Smart Query Definition for Content-Based Search in Large Sets of Graphs

Abstract

Graphs are used in various application areas such as chemical, social or shareholder network analysis. Finding relevant graphs in large graph databases is thereby an important problem. Such search starts with the definition of the query object. Defining the query graph quickly and effectively so that it matches meaningful data in the database is difficult. In this paper, we introduce a system, which guides the user through the process of query graph building. We propose three approaches for graph definition. First, query by example selection starting from an overview of the graph types in the database, second query by sketch combining graph building blocks (i.e., topologic subgraphs) with free graph drawing, and third a combination of both approaches. In all three query definition ways, we support the user with intelligent, data dependent recommendations. It covers the whole spectrum of building parameters such as representative examples, frequent building blocks, or common graph size.