Hidden Group Detection Based On Evolution Of Hidden Markov Model

With the rapid development of information technology, networks have a huge amount of information every day, this makes some of the illegal groups easier to hide its own information.We call these communities hidden groups, which may try to hide its existence and function.Over the past, to identify hidden groups based on hidden markov model, we usually assume that some members consititute a hidden group first, then validate our assumptions on the model. However, this method is not only complicated but also have an enormous amount of calculation, so on larger network it could do nothing at it.Aiming at it, this thesis presents a method for generating communications based on evolution of energe and grouping membership. In this way, we do not make an assumption on the model any more, but start from the communications, completed the detection of hidden group positively. At last, we present the results of experiments on synthetic data as well as real communities, e.g,Enron email to show that the algorithm is effective and feasible.