OLAP on Structurally Significant Data in Graphs

TL;DR

This paper introduces a measure called Structural Significance to identify important data in graph analysis for OLAP, and proposes an algorithm, iGraphCubing, to efficiently compute the graph cube focusing on structurally significant data.

Contribution

It presents a novel Structural Significance measure and an algorithm for structure-aware graph cube computation, improving OLAP analysis on large, high-dimensional graphs.

Findings

Encouraging results on real and synthetic datasets

Effective identification of structurally significant data

Improved efficiency in graph cube computation

Abstract

Summarized data analysis of graphs using OLAP (Online Analytical Processing) is very popular these days. However due to high dimensionality and large size, it is not easy to decide which data should be aggregated for OLAP analysis. Though iceberg cubing is useful, but it is unaware of the significance of dimensional values with respect to the structure of the graph. In this paper, we propose a Structural Significance, SS, measure to identify the structurally significant dimensional values in each dimension. This leads to structure aware pruning. We then propose an algorithm, iGraphCubing, to compute the graph cube to analyze the structurally significant data using the proposed measure. We evaluated the proposed ideas on real and synthetic data sets and observed very encouraging results.