SLS method for the frequent subgraph mining problem (FSM)

Abstract

Graph mining is an important research area within the domain of data mining. The field of study concentrates on the identification of frequent subgraphs within graph data sets. This paper proposes a stochastic local search (SLS) meta-heuristic to solve the Frequent Subgraph Mining problem (FSM). We introduce the notion of diversification which consists on exploring new neighbor solutions that constitute one of the most successful and widely used approaches for solving hard combinatorial problems. A mined subgraph is defined by an objective function which uses two parameters, support and size. The maximization of the size parameter directs the search for finding large subgraphs. The proposed method was implemented and evaluated on two synthetic and five real-world datasets of various sizes and compared to the Local Search (LS), Genetic Algorithm (GA) and Variable Neighborhood Search (VNS) algorithms proposed in the literature. The proposed method is able to discover efficiently diversified subgraphs in the search space by exploring new solutions through the use of randomness in the search. The numerical results show that in general SLS method provide competitive results and finds high quality solutions compared to LS, GA and VNS.