Research On The Communication Structure Of Codeplex Open Source Community Using RMT

The open source community is not only the open source producer production and innovation platform,but also a necessary way to communicate with each other.Over time,the number of open source community projects are increasing,the number of people involved in the discussion is also changing,and these acts are spontaneous,so there will be people enter and exit,resulting in the continuous changes in the structure of communication network.The open source community research mainly is based on the theory of social exchange theory,software development,etc.,which has certain limitations.Therefore,the open source community needs new theories and methods.The random matrix theory(RMT)is a unique visual angle and successful application in practical physics to provide a good opportunity for the development of complex networks.The communication network of the open source community can express the structural information in a matrix.This paper takes the network structures of Codeplex open source community C# from 2012 to 2007 as the research object,constructing the Boolean matrix,the weight matrix and the correlation matrix,and matches the characteristic value sequences with the energy spectrum of the random matrix theory which are obtained through the communication network matrix eigenvalues.Then under the random matrix theory framework,,we can analyze the behavior characteristics of the whole interaction with actual datas from the characteristic spectral density and localization properties and spectral sequence gap and any other aspects.The characteristics of spectral spectrum density analysis showed that the Boolean matrix and weight matrix have a peak at zero eigenvalues.Since Boolean matrix and weight matrix are large sparse matrixes,there has a trailing phenomenon in the biggest eigenvalue.While,the correlation matrix doesn't have these characteristics.This is determined by the characteristics of the internal structure of the matrix,which is characterized by the details of the network.Local characteristics of the network study had found that communication network system has been constantly fluctuating between the state of complete localization and complete diffusion,and closes to the complete localization.Most of the time,the nearest neighbor level spacing distribution obeys the Brody distribution,only in certain periods the network tends to Poisson distribution,which means the levels of the network are almost not related sometimes.These results hint that certain analogies may exist between the spatial topology of complex networks and the temporal evolution properties of quantum dynamical systems.In a word,this paper studies the structure of the open source community from the new perspective of random matrix theory,to help people understand the information communication structures of the production community in the Internet Era.This is also the actual research of mass production,which has certain significance to the development of Chinese opening source community.The research results have strong economic management significance for future knowledge production.