Fig. 2

Frequent subgraph mining. Before starting with the mining, input graph data needs to be properly encoded. After that, the first step towards finding frequent subgraphs is generating a set of candidate subgraphs. Then, the frequency of each subgraph in graph dataset will be checked. This is usually preceded by the pruning of the search space and removal of the redundant candidates with the goal of reducing the search space. It is typically an iterative process: larger candidate subgraphs are generated from smaller frequent subgraphs. The counting step outputs the occurrence of the each subgraph that has been checked and this information is used to calculate the subgraph’s interestingness