Research On Chinese Marking Network Based On Complex Network Theory

In the21st century, the rapid development of complex network has made people pay much attention to the complexity of the network structure and the relationship between network behavior. At the same time, studies of the structure and function of complex networks has become an active topic in statistical physics and its related interdisciplinary subjects. With a quite complex structure, language is also one type of complex network, which means that it is difficult to use traditional linguistics research methods to study the overall characteristics of language. In order to better understand the topology structure and the formation characteristics of language network, this thesis employs a method based on complex network theory to study Chinese language relationship word network, and uses this method to analyze some typical statistical characteristics of the network.This article is organized as follows. Chapter1introduces the significance of complex network research and the domestic and foreign research status of the network of language. Chapter2introduces the related knowledge needed in this article. First describe the theory of random graph, Euler problem of the famous seven Bridges as well as the properties of the ER random graph; Secondly, introduce an important statistical characteristics of the complicated network, the small-world effect, including the most famous of Six Degrees of Separation (Six Degrees of Separation); Finally introduce three statistical properties about the complex network structure, including average path length, clustering coefficient and degree distribution. Chapter3mainly introduces the method to construct Chinese relationship tag network and its statistical properties. After introducing some existing typical topology related research, this paper imports the network database, the457relationship marker words, and then we construct a Chinese relationship network according to database construction. Finally, combined with knowledge about the complex network mentioned before, and with the aid of Pajek software, we investigate the statistical properties of this network. The experimental results show that the relationship marker network has two basic properties of complex networks:(1) small world effect;(2) the network relationship between word degrees generally presents the power-law distribution, words relationship network with scale-free characteristics. Chapter4summaries the work of this paper, including the significance of Chinese relationship tag network research, its theoretical value and some further research directions.