Inference network building and movements prediction based on analysis of induced dependencies
State prediction of a node is made through analyses of adjacent nodes to learn about movements and to manage users in inference network. This study first simplifies ego network to reduce the complexity of structural analyses based on three different relationships among users. Second, degrees of mutual influences are computed according to different forms of association. Finally, the action of the centre node is predicted by analysing the changing of the states of adjacent nodes. The experiment proves that accuracy and efficiency of the algorithm proposed performs well not only in the simplification of ego network but also in the prediction of movements.