A Neighborhood-preserving Graph Summarization

TL;DR

This paper presents a graph summarization method that reduces large graphs by preserving neighborhood structures, enabling faster analysis while controlling information loss, demonstrated through experiments on various graph algorithms.

Contribution

The paper introduces a neighborhood-preserving graph summarization technique that allows adjustable compression for large graphs, balancing size reduction and information retention.

Findings

Significant graph size reduction achieved.

Trade-offs between speed-up and accuracy demonstrated.

Effective for node embedding, classification, and shortest path tasks.

Abstract

We introduce in this paper a new summarization method for large graphs. Our summarization approach retains only a user-specified proportion of the neighbors of each node in the graph. Our main aim is to simplify large graphs so that they can be analyzed and processed effectively while preserving as many of the node neighborhood properties as possible. Since many graph algorithms are based on the neighborhood information available for each node, the idea is to produce a smaller graph which can be used to allow these algorithms to handle large graphs and run faster while providing good approximations. Moreover, our compression allows users to control the size of the compressed graph by adjusting the amount of information loss that can be tolerated. The experiments conducted on various real and synthetic graphs show that our compression reduces considerably the size of the graphs. Moreover, we conducted several experiments on the obtained summaries using various graph algorithms and applications, such as node embedding, graph classification and shortest path approximations. The obtained results show interesting trade-offs between the algorithms runtime speed-up and the precision loss.