Mining Partially Labeled Subgraphs Using Support Constraints

In general, patterns that contain only a few subgraphs will tend to be interesting if they have a high support, whereas larger patterns can still be interesting even if their support is relatively small. A better solution lies in exploiting support constraints, which specify how weakening support is required for what subgraphs, so that only the necessary subgraphs are generated. In this paper, a frame work of frequent partially labeled subgraphs mining is presented in the presence of support constraints, which is based on pattern weakening. This approach is to push forward support constraints into the process of mining so that the proper support is determined for larger subgraphs at runtime to preserve the essence of mining result. The performance of algorithm is evaluated in a multi-group synthetic datasets, and the experimental results show that the method is efficient and fast.