Research On The Complexity Of Mobile Social Network Based On Sina Weibo

Mobilization of traditional social network is facilitated by the birth of iPhone in a large scale. Nowadays, mobile social network has been coming an indispensable component of human’s daily life. High quality mobile social network like Twitter and Path can accumulate plethora users in a relative short time, meanwhile traditional SNS such as Facebook and renren is speeding up its deployment on mobile device. The largest mobile social network in Mainland China is Sina Weibo that holds over350million users to the end of Jun.2012, of which more than60%access the service mainly on mobile device.Considering the similarities between mobile social network and complex network model, construct complex network model using mobile social network subscribers as nodes, relationships between friends as directed arcs. In this article, Sina Weibo were used as research object, an algorithm using Sina Weibo Open API is constructed to overcome the illegal usage of web crawler which is under the limitation of API and ensures the memory deallocation. In the end, a sample of Sina Weibo user network that has2999nodes with232623edges has been collected as an example of mobile social network under the restraints of Sina Weibo open API. Three aspects of study has been made with the data:In the aspect of structure feature of complex network, through the calculation, sample data has an average in-degree of77.57with its longest path length6, which has verified the Six Degrees of Separation in the Sina Weibo user network. According the sample data, in-degree is power law distributed which applies the scale-free network, while out-degree exponent distributed which applies the ER random network. The direct clustering coefficient is generally larger than the inverse clustering coefficient with an average both0.2, which indicates that the network is somehow dense. Most of betweenness centrality is relatively low, while some nodes achieves over1. In comparison with Twitter, less network density and controllability of node’s centrality is found with Sina Weibo network.In the aspect of robustness of complex network, the general approach to measure the robustness is simulating the random failure or the intentional attack. Both random failure and intentional attack has been simulated on the network of sample data, with a brand new and more practical way called random targeted attack. The random targeted attack destroys most important node in probability, important nodes would change its status to become a normal node and would not be deleted from the network immediately until the second attack. The random targeted attack is more suitable to depict the practical situation when a node is under attack. A surprising robustness of Sina Weibo user network has be unveiled.In the aspect of user group discovering of complex network, two algorithm of network communityy discovering have been introduced-hierarchical clustering and GN algorithm. In the later research, hierarchical clustering approach has been deployed to clustering all2999nodes, two charts of network structure have been plotted, but no obvious community has been found in the sample data network of Sina Weibo.Last but not the least, several pieces of advise have been proposed to enhance the data mining and complicity discovering the mobile social network.