FS³: A sampling based method for top‐k frequent subgraph mining

Abstract

Abstract-Mining labeled subgraph is a popular research task in data mining because of its potential application in many different scientific domains. All the existing methods for this task explicitly or implicitly solve the subgraph isomorphism task which is computationally expensive, so they suffer from the lack of scalability problem when the graphs in the input database are large. In this work, we propose FS 3 , which is a sampling based method. It mines a small collection of subgraphs that are most frequent in the probabilistic sense. FS 3 performs a Markov Chain Monte Carlo (MCMC) sampling over the space of a fixed-size subgraphs such that the potentially frequent subgraphs are sampled more often. Besides, FS 3 is equipped with an innovative queue manager. It stores the sampled subgraph in a finite queue over the course of mining in such a manner that the top-k positions in the queue contain the most frequent subgraphs. Our experiments on database of large graphs show that FS 3 is efficient, and it obtains subgraphs that are the most frequent amongst the subgraphs of a given size.

A survey of pattern mining in dynamic graphs

“…Although frequent subgraph mining is useful, a problem is how to set the minsup threshold (Duong, Khan, Jeong, & Lee, 2016;Fournier-Viger, Cheng, Lin, Yun, & Kiran, 2019;Li, Lin, Li, & Duan, 2010;Saha & Hasan, 2014). If it is set too high, few patterns are found, while if it is set too low, too many patterns may be found and algorithms may have very long runtimes or run out of memory.…”

“…Hence, the algorithm may return infrequent patterns and miss frequent patterns. Another approximate algorithm named kSIM (Saha & Hasan, 2014) was proposed to process a restricted type of graphs called induced subgraphs, and was shown to outperform FS 3 in runtime and accuracy. Recently, an exact algorithm for top-k frequent subgraph mining was presented, named TKG (Fournier-Viger, Cheng, Lin, Yun, & Kiran, 2019), which extends gSpan and has similar performance.…”

Mining exceptional closed patterns in attributed graphs

“…It takes benefit from a closure operator to avoid redundancy and efficiently prune the search space. The second algorithm mines closed exceptional subgraphs by directly sampling the space of closed patterns in a similar way as [3,11,28].…”

“…Also, this approach has not been applied to richer pattern languages yet. Some researchers have tackled the problem of sampling the output space of frequent subgraphs in a collection of graphs [11,28]. These methods are based on random walks.…”

“…These methods are based on random walks. In particular, [28] aims at returning the top k frequent graphs of a specified size. The problem we tackle is different on several points: We consider a single graph and a discriminative measure instead of a frequency measure.…”