Information Propagation Model Based On Opportunity,Trust And Motivation And Prediction Analysis

The rapid development of large-scale online social network sites,such as Facebook,Twitter and Microblog,indicates that social network has become a key platform for people to obtain information,publish information and social interaction.A large number of node users actively participate in the process of information generation and propagation,which makes the information presented with the characteristics of comprehensive and rich content,rapid and accurate dissemination,wide range of dissemination and deep influence.An excellent model of information dissemination and its prediction evolution can make people have more accurate and intuitive understanding of the communication methods such as public opinion,marketing,which is conducive to discovering the weak links of the modes of communication,and then predict the degree of potential threaten,and formulate relevant control strategies.Especially,it has a clear positive social significance to effectively control and guide the network public opinion which threatens social security and causes social unrest.However,most of the current information dissemination and its prediction evolution models exist two problems: one is that they take information content and network topology as the dissemination conditions,and do not take into account the social or emotional factors of the node users in the network;the other is that the existing models mainly consider the static attributes of users,while ignoring the situation that users' behavior or interests will change with time,and do not have the ability to predict information dissemination.In order to solve the above problems,the dissemination and evolution prediction of information in microblog social network is studied in this thesis.The main research contents include the following two aspects.(1)Based on the principle of ternary closure in social belonging network,a hybrid social factors information dissemination model based on opportunity,trust and motivation is proposed.Firstly,the interest similarity between two users is convenient for measuring the opportunity to receive certain information.Secondly,the threshold of social trust in the process of information dissemination is determined by coupling users' network influence and content contribution.Then,considering that each user in the network always chooses the information of the best benefits to spread,the game evolution theory is used as the motivation for users to spread a message.Finally,the game choice information propagation model based on page rank algorithm(GCIP-PageRank model)is proposed.The experimental results show that the proposed method has achieved good results.(2)This thesis proposes a prediction model of information propagation based on time series characteristics(TSPIP model),which synthesizes the structural characteristics ofmicroblog community and the data with time series.Firstly,the time series of microblog hot events(including stationary and non-stationary events)are detected,and the non-stationary event propagation sequence is transformed into event propagation sequences and stationary event propagation sequences.Secondly,a linear regression model is used to establish the quantitative relationship between the quantity of data and the explanatory variables,and the best explanatory variables are selected by the correlation coefficient and the hypothesis test.Finally,the explanatory variables and associated data are introduced into the model.The state space model of Local Linear Trend and Regression is used to construct the predictive model of microblog hot events propagation,and the time series analysis is conducted to obtain the characteristics of event development trend and cycle.Experiments on real data sets show that our prediction method is effective for the propagation of microblog hot events with burst traffic.