Locating the Social Network Suspects by Using Subsequence Matching and Graph Clustering

초록

In this paper, one of the social network relationship issues such as Stalker or Paparazzi for the user trajectory as well as social network activity data is explored by using the Subsequence Matching and Graph Clustering. It has frequently been applied the K-means clustering method in many clustering environment. We exploit a Subsequence Matching and Graph Clustering method, because each and every point in the social network domain constitutes trajectories as a graph rather than individual points. The similarity for the subsequence matching and Graph Clustering should have been devised for the route on sequence data based on the number of contact for mobile network. Experimental results represent that our approach effectively detect the suspect for the various synthetic data sets.