Research Of Copyright Content Propagation Based On Pirates Community Discovery In Peer To Peer Networks

Due to its decentralized nature, self-organization, and dynamic scalability, etc., P2P network has been applied in resource sharing and content propagation widely. Unfortunately, the popularity of P2P network suffers from a rampant piracy propagation problem for lacking of strict content authorization and third party supervision, is facing a serious crisis of confidence. In order to excavate the sharing relationship of copyrighted contents among users, reflect the process of contents transmission and distribution in peer-to-peer network deeply, and ultimately regulate the copyright contents and prevent the propagation of the piracy content in an effective way, this paper introduced the theory of community discovery and proposed a community discovery approach and propagation system based on actor-network model. In the system, community structure is as the object of propagation. Based on the research of the distribution of content in the community, the community behavior threat degree is quantified and the strategies to counter the dissemination of infringing content community are explored further.User behavior-based method has been proposed, that the user node which has higher threat degree is found and removed in the network. Content similarity graph-based method also argued that we can calculate the content propagation ability to implement the analysis and regulation of copyright content. Pirates distributed and reproduced copyrighted contents or illegal pornography without authorization. When pirates frequently share similar piracy contents with each other, they naturally form the pirate community. Community discovery technology is one of an important research in data mining and exploratory analysis technology in complex network research.In this paper, based on the actor-network model, this research method constructs pirate community using the following steps. Firstly, network data preprocessing. Secondly, actor network model building. Thirdly, pirates community discovery. Step two includes the sub-steps:copyright content similarity graph structuring, nodes relationship connecting and message filtering that simplifying the mess P2P network. In the copyright content similarity graph structuring step, to make sure the accuracy of copyright content description and comprehensive mining, we use ontology to descript the content metadata firstly and calculate the similarity. In nodes relationship connecting step, a message weight is calculated based on message time stamp, message passing frequency and massage type. In the third step, we design a divisive algorithm ANMGN (ANM-based GN algorithm) that can remove more edges in one recalculation based on GN algorithm. Although this algorithm has been applied successfully to community analysis in a variety of networks giving the modularity to quantify the strength of community structure, it does not offer a strategy to break the recalculation loop. Here we apply the increment of modularity to break the loop which will reduce the runtime.So far, one method to copyright content supervision is in a post processing way, that is, once piracy is detected, the system will prevent the spread of infringing content by stop download or file removing. Another way is encryption technology by content format but it is private by specific software developer. In order to achieve regional early warning, we analysis the community threat degree and propose two kinds of counter mechanism:node-broken strategy and edge-broken strategy. Through removing the contact node or pivotal edge, we can block the propagation of copyright content.