Multi-level Graph Drawing Using Infomap Clustering

Abstract

Infomap clustering finds the community structures that minimize the expected description length of a random walk trajectory; algorithms for infomap clustering run fast in practice for large graphs. In this paper we leverage the effectiveness of Infomap clustering combined with the multi-level graph drawing paradigm. Experiments show that our new Infomap based multi-level algorithm produces good visualization of large and complex networks, with significant improvement in quality metrics.