Research On Information Network Model Based On Graph Theory

Graph structure, a widely used structure, which has the characteristic to demonstrate the relationship between several nodes, is the clear visual expression of the real world. As the theoretical basis of the graph structure at present, graph theory is widely used in social networks, genetic computation and social patterns. What’s more, it has lead trend of the development in its field due to the outstanding performance and direct manifestation.The information network studied by this thesis is the abstract expression of graph structure to the real world based on graph theory. It’s also a calculation mode based on this data structure. As the world’s newly developed field, graph computing has better performance and direct manifestation in social networks, genetic computation and social patterns when compared with traditional matrix calculation. However, because of the research of graph computing are relatively new, recent researches are focusing on the relationship study and other aspects are still on the blank stage. According to the basic theory of graph computing and the needs of reality, this thesis has studied the structure of data processing, sub graph filtering and query, association and correlation, which used to solve the problems of graph computing.Two aspects are consisted in the research of this thesis:One of them is the theoretical research based on graph theory, which provided theoretical foundation for model construction. Another is the practical technical research based on natural language processing, which provided analytical and constructive tool for model.The proliferation of computers brings the development of the Internet, lots of information and messages stored in the form of electronic documents. As the user of Internet, data collection and difficulties of use should be considered when facing vast amounts of information. When facing great demand and challenge brought by vast amount of information, people need the tool, which can acquire real data from several information automatically.Thesis contains five parts:the first part includes the trend of the development of Internet today and the recent situation of graph theory in the field of Internet; the second part is the definition and description of graph theory; the third part introduces the construction plan of network model based on graph theory; the forth part is the research of related arithmetic of information network model, including follow-up of hot issues, technology of entity similarity information prediction, calculation of node similarity and node clustering, etc.; the fifth is the conclusion and perspective of this thesis.During the research of this thesis, we have combined the recent trend of development in Internet by using advanced technologies like nature language processing and entity extraction. What’s more, we also combine these theories with graph theory so that they can be inseparably interconnected. The research of this thesis not only solves the problem of constructing and storing information network model, but also deeply studies the property and arithmetic of information network model. In the basis of the information network model, it solves several aspects including follow-up of hot issues, node clustering, and logical deduction, accomplishes the anticipative aim to information network model.