Estimating Edge-Local Triangle Count Heavy Hitters in Edge-Linear Time and Almost-Vertex-Linear Space
2018 IEEE High Performance extreme Computing Conference (HPEC)(2018)
摘要
We describe a methodology for estimating edge-local triangle counts using cardinality approximation sketches. While the approach does not guarantee relative error bounds, we will show that it preserves triangle count heavy hitters - the edges incident upon the largest number of triangles - well in practice. Furthermore, we provide empirical evidence that the sum of edge-local estimations yield reasonable estimates of the global triangle count for free. In this paper we describe a two-pass algorithm for estimating edge-local triangle count heavy hitters. The algorithm requires time linear in the number of edges, memory almost linear in the number of vertices, and is easy to parallelize. We provide results on dozens of real-world and synthetic graphs.
更多查看译文
关键词
graphs,triangle counting,sublinear algorithms,estimation algorithms
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络