Counting Butterfies from a Large Bipartite Graph Stream

Abstract

<p>We consider the estimation of properties on massive bipartite graph streams, where each edge represents a connection between entities in two different partitions. We present sublinear-space one-pass algorithms for accurately estimating the number of butterflies in the graph stream. Our estimates have provable guarantees on their quality, and experiments show promising tradeoffs between space and accuracy. We also present extensions to sliding windows. While there are many works on counting subgraphs within unipartite graph streams, our work seems to be one of the few to effectively handle bipartite graph streams.</p>