Communities of Interest

摘要: We consider problems that can be characterized by large dynamic graphs. Communication networks provide the prototypical example of such where nodes in graph are network IDs and edges represent communication between pairs IDs. In graphs, appear disappear through time so methods apply to static graphs not sufficient. introduce a data structure captures, an approximate sense, its evolution time. The arises from bottom-up representation as union small subgraphs, called Communities Interest (COI), centered on every node. These subgraphs interesting their own right we discuss two applications area telecommunications fraud detection help motivate ideas.