Data Grab And Empirical Analysis Of SNS-oriented Social Network

With the rapid development of the World Wide Web, the social network in cyber world expands fastly and its structure is so complex that the traditional methods like figures, tables or text are useless to analysis, show and manage massive data of no treatment, making a wealth of information resources contained hard to be discovered and used effectively. This thesis mainly focuses on structural investigation from the photo-sharing website Flickr and crawls a social network dataset from it and analysises the out-in degree of dataset and compreses the dataset through SQL programs on SqlServer 2005 platform, the local social network surrounding a given user (noted as the central node in the local network) is extracted, some key nodes with big degree in local social network are selected, the nodes locating on the shortest path of the central nodes are selected, and eventually three key social networks are retrieved and represented in matrix forms, an algorithm is designed and implemented to initial layout nodes in the key social networks with a "concentric circles" model,the thesis also designs and implements algorithms to layout and display the key network on Matlab platform.The social network Flickr has scale-free's characteristic of complex social network through the study of social network datasets, the thesis puts forward a compression scheme through SQL programs and it improves the compression efficiency of the datasets.