The Research And Application Of Complex Network Modeling

At the end of last century, with the development of information technology such as the Internet, research of complex network gradually aroused people’s attention. Network has its own diversity and disruptive making it become one of the most tremendous challenges in complex system research. In recent years, the small world model and scale-free network model based on statistical physics of complex networks makes the complex network research scholars upsurge again. How to put forward a reasonable model to describe the actual network system and utilize complex network modeling methods to analyze the problems in the real world to reveal its law are the most critical two questions.At present, the classic model in complex network science can simulate the real part of the complex network system, however, not conform to the characteristics of real network. Edge connection and node adding rules for the scale-free network model are increasing at a constant speed, against the actual situation increasing at a variable speed. Therefore, proposing a model increasing at a variable speed of the complex network model, and applying the existing complex network model and its topological characteristics in the actual work are of great significance.This paper focuses on the study of complex network modeling method and employs the research methods for complex network to analyze the real network. The main works is as follows:1. Proposition of two new network models. According to the characteristics of the nonlinear increasing in the real world network, we put forward two kinds of new network model based on the different common renewal counting processes, simulation experiments and calculation the index of the model by using mathematical method can prove that the presented two kinds of model are in accordance with the characteristics of complex networks.2. Modeling of the practical power transaction network. Apply the existing network model to a practical application of the real world: power transaction process, and obtain the electric trading network. Then, we use the complex network knowledge to analyze the electric network and conclude the characteristics of network by simulation experiments.3. Application of the complex network theory on natural language and proposition of a new keyword extraction method based on complex network. The adjacent relation between the words in the text are on the ground of modeling complex network. By analyzing the topology of the network, we find out that using node betweenness and node degree as the basis of keyword extraction will improve the accuracy of the keyword extraction.