Research On Modeling And Algorithms Of Human Dynamic In Online Social Networks

The exploration of the law of human behavior becomes very important and interesting due to its complexity. It is also because human behavior is highly responsible, it has been attracting research interest from sociology, psychology, economics and other disciplines of scholars to discover and reveal the objective law. In recent years, with the rapid development of network technology, online personal communities, online dating networks and online social media, as a representative of online social network are widely used. Especially with the development of mobile communication technology, people can enjoy a variety of social application conveniently, the development of application directly brings the increase of information, a lot of data of user behavior is recorded in real time, the age of big data gives us a hitherto unknown opportunity to study human behavior.In this paper, we research and explore the user behavior at the level of individual memory and the level of interactions among users in social network. First of all, we research the dynamics that drives user behavior, and put forward the user behavior model from the above two aspects. Then, based on the model of research findings and results, we discuss the algorithm of discovering a role and topic detection as well as tracking method of evolution. Finally, the author develops an analysis, guidance and control system of the behavior of online social network. In detail, the main innovations of this paper are mainly reflected in the following5aspects:1. The first step studies the human behavior in the top-one social network system in China. By analyzing real-life data at a large scale, we find that the message releasing interval (intermessage time) obeys power law distribution both at individual level and at group level. Statistical analysis also reveals that human behavior in social network is mainly driven by four basic elements:social pressure, social identity, social participation, and social relation between individuals. Empirical results present the four elements’impact on the human behavior and the relation between these elements. To further understand the mechanism of such dynamic phenomena, a hybrid human dynamic model which combines "interest" of individual and "interaction" among people is introduced, incorporating the four elements simultaneously. To provide a solid evaluation, we simulate both two-agent and multiagent interactions with real-life social network topology. We achieve the consistent results between empirical studies and the simulations. The model can provide a good understanding of human dynamics in social network.2. The increasing development of social networks provides a unique source for analyzing human dynamics in the modern age. We analyze the top-one Internet forum in China and identify the statistical properties of hotspots, which can promptly reflect the crowd events in people’s real-life. Empirical observations indicate that the interhotspot distribution follows a power law. To further understand the mechanism of such dynamic phenomena, we propose a hybrid human dynamic model that combines "memory" of individual and "interaction" among people. To build a rich simulation and evaluate this hybrid model, we apply three different network datasets. Our simulation results are consistent with the empirical studies, which indicate that the model can provide a good understanding of the dynamic mechanism of crowd events using such social networking data. We additionally analyze the sensitivity of model parameters and find the optimal model settings.3. Mastering human behavior plays an important part in user classification, network marketing and public opinion tracking etc. Different groups of users have different behavior characteristics. Thus, it is very important and interesting to discover human social role based on their behavior. The rise of social networks, and massive user behavior have been recorded, provide us great potential to do this exploration. The present user behavior characterization methods are generally based on the structural features or static coarse-grained behavior features of nodes. In this paper, by analyzing a large number of human behavior data, we propose a new methodology for social role discovery based on dynamic and fine-grained human behavior attributes in social networks:We first mining and propose number of properties that on behalf of human behavior; After that, to deal with the large human behavior data, a simple, scalable and distributed parallel clustering algorithm based on grid and density is developed; The theoretical analysis and experimental results show that the algorithm has better efficiency and effectiveness, and algorithms reveals valuable discovery on the real-life datasets;Besides, the methodology in this paper for user role discovery also can be applied to social networks in general.4. On the basis of analyzing the characteristics of user behavior on the group level, we research the user roles and different roles concern of hot topic in online social network. Considering the features of a new generation of online social network information in Micro-blog system as the representative of the brief, semantic sparse, we take the customer as the center, according to the user behavior, overlay network of keywords and user relationship is proposed based on the implicit links. And also we excavate the important user and cluster the important user based on relation network, which indicates that different types of user groups of function in the life cycle of the hot topic. In addition, it can also be useful for nature, understanding of subject characteristics by finding out the important users in a different dominant topic from a variety of hot topics and using the relationship between the users.5. According to the above research, we develop an analysis, guide and control system of social network based on user behavior of Sina Micro-blog-SayToYou. Because of the characteristics of huge amounts of data, such as fragmentation and loose, based on the theory of complex networks, from the multiple dimensions of time, space, count, role to analyse the fine-grained user behavior in different levels. Based the user behavior referred to this paper, we achieve a hierarchical social network of user behavior of analysis, guidance and control system to address user behavior in large data environment, public opinion tracking and a hot topic of centralized, unified visualization of effective management issues, and to achieve the purpose of online user behavior guide and auxiliary control. At the same time, the system also has a feedback to the offline actual social network.In a word, the paper discusses the models, algorithms and practice of user behavior in online social networks in the individual and group levels from the theory and engineering aspects respectively.