Efficient Algorithms for Approximate Triangle Counting

Abstract

Counting the number of triangles in a graph has many important applications in network analysis. Several frequently computed metrics like the clustering coefficient and the transitivity ratio need to count the number of triangles in the network. Furthermore, triangles are one of the most important graph classes considered in network mining. In this paper, we present a new randomized algorithm for approximate triangle counting. The algorithm can be adopted with different sampling methods and give effective triangle counting methods. In particular, we present two sampling methods, called the \textit{$q$-optimal sampling} and the \textit{edge sampling}, which respectively give $O(sm)$ and $O(sn)$ time algorithms with nice error bounds ($m$ and $n$ are respectively the number of edges and vertices in the graph and $s$ is the number of samples). Among others, we show, for example, that if an upper bound $\widetilde{Δ^e}$ is known for the number of triangles incident to every edge, the proposed method provides an $1\pm ε$ approximation which runs in $O( \frac{\widetilde{Δ^e} n \log n}{\widehat{Δ^e} ε^2} )$ time, where $\widehat{Δ^e}$ is the average number of triangles incident to an edge. Finally we show that the algorithm can be adopted with streams. Then it, for example, will perform 2 passes over the data (if the size of the graph is known, otherwise it needs 3 passes) and will use $O(sn)$ space.