Fractal Application And Evolution Researches Of Mobile Social Networks

Mobile social networks(MSNs)are a kind of social networks formed by collecting and clustering the location information of the smart mobile devices.They are formed by the internet behavior of many social groups which consist of more and more people using mobile devices such as smart phones.MSNs are designed to make communication between people more convenient and faster,and they greatly reduce the cost of information dissemination and exchange.They can also reduce the time of information collecting,and provide better social services for people of all ages and all works around the worldThis thesis focused on fractal applications and evolution laws of MSNs.In the first experiment of this thesis,it used the clustering coefficient as measure,and it was found that the time series of MSNs have fractal properties by using rescaled range analysis(R/S analysis).In the second experiment,it used the superfamily analysis of networks to prove that superfamily properties did exist in one single MSN.Changes in MSNs and network predictions can be found and detected by using the conclusions.(1).In the part of fractal application of MSNs,R/S analysis was used to analysis the clustering coefficient time series of MSNs.The results also confirm that the clustering coefficients time series of MSNs is fractal.Firstly,we converted the original network to a weighted network.Secondly,by fitting the Range and the standard deviation of clustering coefficient,it proved that MSNs have fractal properties in dimension of clustering coefficient aspect.Finally,by analyzing the Hurst and ?Hurst exponent,it is found a relationship between the sudden change of clustering coefficient time series and Hurst exponent and ?Hurst exponent,and we can use this finding to make some simple predictions.(2).In the work of evolution of MSNs,it cut the weighted MSNs in fixed length of time interval.It used the superfamily analysis to analyze these time slices,and it was found that superfamily property did exist in one single MSN.Then by superimposing the continuously time intervals,it captured a change in the network.The SP values of the superimposed time intervals had changed in a certain period of time.This approach demonstrated the evolution of MSNs with time went on.