Extracting top-K interesting subgraphs with weighted query semantics

Heterogeneous information networks (HIN) contain abundant of information about entities i.e. people, places, organizations, and events etc. with their relationship. Extracting interesting information from such underlying networks is important in real world which essence can be related to subgraph search problem. Previous approaches focuses only on structural matching with naive ranking by summing up edge weights of a subgraph in the HIN. To this end, we have proposed concept of weighted query measures. Specifically, we propose two types of interestingness measurements based on weighted query semantics, Influential Edge Match (IE-Match) and Closest Match (C-Match). Moreover, we devise Destination Vertex Count (DVC), a space efficient indexing scheme to improve indexing and candidate generation process. To evaluate and show effectiveness of the proposed approach, we conduct extensive experiments on synthetic and real world datasets and present interesting real world case studies.