Complexity Research On Social Media Entity Association Networks

With the rapid development of Web2.0, blogs, social networks, micro-blog andother social media have emerged and gradually prevailed. Internet as a social mediahas become an important part of people’s daily life and sociality platform. Complexnetworks research is a state-of-the-art hot topic, which has made substantial progressrecently. Complexity theory provides us a new scientific point of view and method toknow and explore the real world. It is a fresh attempt to study on the complexity ofentity association networks in social media. The main content of this study includes:entities extraction in social media; building complex networks based on the entityrelationships; making deep research on it, in order to achieve a more in-depthunderstanding of social media entity network.Firstly, the foundations of social media and complex networks are introduced,and the research paradigm based on complex network in social media is reviewed.Secondly, a knowledge network of the relationships between characters is constructed.The relationships are extracted from Wikipedia. Complex networks characteristics ofthe network show that it is a small world and scale-free network, which is consistentwith real social networks and similar to other social networks. In order to investigatethe kernel nodes of the network, centrality analysis is performed, and the result showsthat, contemporary figures are largely concentrated in entertainment, while man of theancient figures are basically political. This find is in accord with intuition of human,since Wikipedia is compiled by web users freely, and the richness of the content inline with man’s interest. Finally, community detection result on the network makes anin-depth analysis of the characteristics of typical members of the community. Resultsdemonstrate that the members of the same community together can be explained asbelonging to the same family, or engaging in the same occupation, or experiencingcommon historical events. As we can see, the method based on complex networks issignificant to promote a fine-grained research on Wikipedia knowledge network.Thirdly, in-depth analysis of micro-blog topic word network. In recent years,researches for the topic analysis and computation are mainly focused on topicdetection and tracking, studies on form analysis of topic relationships, topic patterndiscovery at macro level and other aspects have not yet been carried out. It is the first time to apply complex network approach to topic pattern discovery at the macro view.Making study on the network topology relation, found that the Weibo topic termnetwork is a typical complex network, and is consistent with the small-world effectand scale-free characteristics of the real social networks, and also similar to othersocial networks. Using multilevel algorithm for community detection, found that thenetwork has obvious community features. At last, presents a micro-blog topic publicopinion hotspot discovery method based on complex network.The hot spots in the network are extracted through centrality analysis methods.C The correlation analysis results for topic extraction can be applied to real life.Complexity research on Weibo topic term network is of great significance for publicopinion analysis and warning.Finally, the conclusion and future work are presented.