Graph cells: Top-k structural-textual aggregated query over information networks

Abstract

The graph OLAP aggregated analysis in information networks has been extensively studied. However, previous works have neglected to integrate the structural information into this kind of query and ignored the influence of enough textual information in graph aggregation operations. In this paper, we propose a novel OLAP query called top-k structural-textual aggregated graph cell query to analyze the information data. According to the given keywords, this query is to find top-k structural-textual aggregated graph cells in text-rich multidimensional information networks. Under the conditions of matching attribution values in a portion of dimensions, a graph cell is defined as a subgraph of the network. It only contains documents of all included vertices in this subgraph. To distinguish the importance of different graph cells, we firstly design a dominating number-based threshold testing and a flexible ranking function integrating the text similarity with the query and the structural size to obtain k most relevant graph cells. Then, we propose a new hybrid index structure and a filtering-and-verification framework, which includes an efficient search algorithm and several pruning and bounding techniques. Finally, we verify the effectiveness and efficiency of the proposed methods through extensive experiments.