HMM-based Connection Between The Social Network Analysis

The online social relationship has been one of the most important social relationships between individuals, growing from computer-mediated interactions. Relationships of social network research and analysis in recent years become a hot spot. As the online social network, the relationship between users is often not visible, but reflected through a lot of interaction, it need to observe the interaction between users to infer or estimate the relationshp between users. In this thesis,HMM-based method to analyze through the implicit connection between the state estimation,that the observed interaction events to observe the state to estimate the intrinsic relationship between the state.This thesis analyzes the social network model,social networking is mainlycomposed of nodes and connections,and connection between the social network user interaction is the result.This thesis describes the relationship between social networks and social relations refined for the connection strength,depth and connection to connect logic.On this basis, the user interaction event with HMM modeling,where the user changes the interaction between the stage even as a Markov,state the relationship between users as hidden Markov state,the two constitute a hidden Markov the model.Then,after learning of HMM hiden state estimation model.In the HMM hidden state estimation of the time,instead of the Viterbi algorithm with the DS,and the two made a comparative analysis, a detailed analysis of both the advantages and disadvantages, and their respective advantages,experiments show that when using a larger data set Viterbi algorithm better.