Social networks represent the social relationships between humans or organizations. Spreading processes on social networks have an extraordinary impact on economy and society. Information diffusion on social networks helps people to obtain information and communicate with each other. But spreading of rumors and harmful information on social networks may undermine the social stability, or even causes crises and emergencies. Besides, spreading of infectious diseases on social networks inflict huge threats not only to economic development but also to the health of mankind. Spreading dynamics on social network has received a great deal of attention in recent years, which aims to understand, predict and control the macro-level characteristics and micro-level behaviors in the spreading process. But, most of previous studies on spreading dynamics on social networks assume homogeneous social relationship, which contradicts the observations of real world. Due to different tightness and multiple types of social relationships, real social systems can be described as weighted social networks and multi-relational social networks. This dissertation studies the spreading dynamics on weighted and multi-relational social networks from three aspects: empirical study, modeling, and control method designing and evaluation. The study has great implications for understanding and controlling the spreading process in the real world. The major works and contributions of this dissertation are as below. (1) An empirical study of spreading dynamics on a weighted social network. Based on the heterogeneous news reprinting transactions between news portals, a weighted news reprinting network is constructed to study the spreading dynamics of online news. The empirical results show that: 1) a small fraction of news portals play important roles on information spreading and collecting, they form a dense connected core, which other news portals reprint news from or to; 2) the infectees’ demands of publishing a large number of news and the infectors’ influence on other news portals are two major factors to determine the number of news articles that reprinted from infectors to infectees; 3) the reports about a news event would spread rapidly through news portals which are interested in the event, and the follow-up reports about the event lead oscillations in the outbreak of the event. (2) Modeling the spreading dynamics on weighted social networks based on individuals characteristics. O...
修改评论