Topological Depiction Method For Complex Network Structures

Big Data revolution is coming to our modern society. It not only brings us the newfeatures of the information, but also changes our thought in a brand new way. Thesechanges deeply affect our ways of doing research. Big Data is often from a complexsystem, and complex network is a brief description of complex system. Topologicaldepiction method for complex network structures is a highlight and difficult problem inthe research field of complex network. Community structures, as one of the importantcomplex network structure, have been paid a lot of attention and many of theachievements have been done.The topological data analysis method is an appropriate tool in coping with themassive data. Barcode depiction method is one of the significant methods above. Thispaper applies the Barcode depiction method in dealing with the research of the complexnetwork structure. And all the work is surrounded by the follow two parts:First, this paper put forward a method of building the complex network of thename entity from the massive text data. It extracts the name entities as the points of thecomplex network, makes up the feature description by using the TF-IDF method andstructures the relationships to obtain the new network by applying the cosine similaritymeasurement. This way of strutting network is able to affect the relationship betweenthe points objectively and significantly.Second, this paper improves the Barcode method as the depiction method of thecomplex network structures. The improved Barcode can effectively descript thecomplex network structures and is quite visible. It can automatically excavate a newkind of community structure: persistent connected structures. Focusing on thesestructures, it precedes the Barcode feature analysis. By proposing and researching theindex of the tightness of these structures and its distribution, the experimentannounces the partitioned power law of the persistent connected structures. This lawaffects the statistics features of the community structures, deepens the knowledge of thecommunity structures and accounts for the availability and theoretical significance ofthe Barcode method.