【摘要】：P2P (Peer-to-Peer)對等網(wǎng)絡(luò)因其非中心化,自組織,動態(tài)可擴展等特點,在資源共享和內(nèi)容傳播方面得到廣泛應(yīng)用的同時,由于缺少一種嚴格的內(nèi)容授權(quán)和第三方監(jiān)管機制,造成了不良資源和盜版侵權(quán)內(nèi)容的泛濫傳播,正面臨著一場嚴重的信任危機。為了深入挖掘用戶節(jié)點之間版權(quán)內(nèi)容共享關(guān)系,深刻反映P2P網(wǎng)絡(luò)中版權(quán)內(nèi)容的傳播與分布,最終有效地進行版權(quán)內(nèi)容的監(jiān)管與侵權(quán)內(nèi)容傳播的反制,本文引入社區(qū)發(fā)現(xiàn)理論,提出基于行為者網(wǎng)絡(luò)模型的侵權(quán)社區(qū)挖掘方法及監(jiān)管系統(tǒng)。以網(wǎng)絡(luò)中社區(qū)結(jié)構(gòu)為監(jiān)管對象,通過對社區(qū)內(nèi)版權(quán)內(nèi)容分布的研究,對社區(qū)的威脅度進行量化分析,并探索侵權(quán)社區(qū)內(nèi)容傳播的反制策略。 通過對對等網(wǎng)絡(luò)中侵權(quán)行為研究發(fā)現(xiàn),侵權(quán)者之間相互共享和傳遞大量相似版權(quán)內(nèi)容,這種基于內(nèi)容共享和消息傳遞的節(jié)點組成社區(qū)的結(jié)構(gòu)。社區(qū)發(fā)現(xiàn)技術(shù)是數(shù)據(jù)挖掘和探測性分析技術(shù)在復(fù)雜網(wǎng)絡(luò)中極為重要的研究,在社會網(wǎng)絡(luò)分析領(lǐng)域、數(shù)據(jù)挖掘和數(shù)據(jù)庫技術(shù)領(lǐng)域、統(tǒng)計學(xué)和機器學(xué)習(xí)等領(lǐng)域有廣泛的應(yīng)用。 本文基于行為者網(wǎng)絡(luò)(Actor-network model)模型構(gòu)建侵權(quán)社區(qū)的研究方法包含：步驟1用于網(wǎng)絡(luò)數(shù)據(jù)的預(yù)處理；步驟2構(gòu)建行為者網(wǎng)絡(luò)模型；步驟3完成侵權(quán)社區(qū)挖掘。其中第二步包含內(nèi)容相似度圖的構(gòu)建、節(jié)點關(guān)系聯(lián)接和消息篩選及過濾等步驟。在內(nèi)容相似度圖構(gòu)建模塊,本文首先對內(nèi)容元數(shù)據(jù)進行本體描述,然后進行相似度的計算,確保版權(quán)內(nèi)容描述的準確性和挖掘的全面性；節(jié)點關(guān)系聯(lián)接模塊,記錄了消息時間戳、消息頻率、消息類型在內(nèi)的性質(zhì)完成消息權(quán)重的量化；消息篩選及過濾過程大大簡化了P2P網(wǎng)絡(luò)數(shù)據(jù),為實施社區(qū)挖掘算法做到提前優(yōu)化。第三步是在研究社區(qū)挖掘經(jīng)典算法Girvan-Newman (GN)算法的基礎(chǔ)上提出有效的侵權(quán)社區(qū)挖掘的ANMGN算法,引入社團結(jié)構(gòu)增益縮短算法時間。 目前,版權(quán)內(nèi)容的監(jiān)管一種是采取事后處理的方式,即發(fā)現(xiàn)盜版侵權(quán)內(nèi)容后,通過采取阻止下載、文件清除等技術(shù)阻斷侵權(quán)內(nèi)容的傳播；另一種是內(nèi)容格式加密技術(shù),符合版權(quán)方許可的用戶才能下載,而這種技術(shù)并不是開源的,只適用于特定軟件,沒有普遍性。本文提出的基于侵權(quán)社區(qū)挖掘方法實現(xiàn)了對版權(quán)內(nèi)容的傳播流向和分布檢測與監(jiān)管,達到區(qū)域預(yù)警及時對侵權(quán)社區(qū)予以打擊。為了解決P2P網(wǎng)絡(luò)中的盜版侵權(quán)內(nèi)容傳播問題,本文在侵權(quán)社區(qū)發(fā)現(xiàn)的基礎(chǔ)上,分析了社區(qū)的威脅度,對侵權(quán)社區(qū)內(nèi)版權(quán)內(nèi)容傳播提出兩種反制機制：斷點反制策略和斷邊反制策略。版權(quán)內(nèi)容傳播時出現(xiàn)經(jīng)過次數(shù)最多的節(jié)點,我們稱之為橋接點,斷點反制策略就是通過找到社區(qū)橋接點,在必要時刪除這些橋接點,以達到阻礙版權(quán)內(nèi)容向外傳播；社區(qū)之間的連接邊,我們稱之為關(guān)鍵邊,斷邊反制策略通過找到網(wǎng)絡(luò)中的關(guān)鍵邊,在必要時產(chǎn)出這些邊,破壞版權(quán)內(nèi)容的傳播。
[Abstract]:P2P (Peer-to-Peer) peer-to-peer network has been widely used in resource sharing and content dissemination because of its decentralized, self-organization, dynamic scalability and other characteristics, at the same time, due to the lack of a strict content authorization and third-party supervision mechanism, As a result of the spread of bad resources and pirated infringing content, it is facing a serious crisis of trust. In order to deeply excavate the relationship of copyright content sharing among user nodes, reflect the dissemination and distribution of copyright content in P2P network, and finally effectively carry out the supervision of copyright content and the counter-system of infringing content dissemination, this paper introduces the theory of community discovery. The method and supervision system of tort community mining based on actor network model are proposed. Taking the community structure in the network as the regulatory object, through the study of the distribution of copyright content in the community, this paper makes a quantitative analysis of the threat of the community, and explores the countermeasures for the dissemination of content in the infringing community. Through the study of infringement in peer-to-peer network, it is found that infringers share and transfer a large number of similar copyright content to each other, which is a community structure based on content sharing and message transmission. Community discovery technology is a very important research on data mining and exploratory analysis technology in complex networks. It has been widely used in the field of social network analysis, data mining and database technology, statistics and machine learning. In this paper, the research methods of building tort community based on actor network (Actor-network model) model include: step 1 for network data preprocessing; step 2 to build actor network model; step 3 to complete infringement community mining. The second step includes the construction of content similarity graph, node relational join, message filtering and so on. In the construction module of content similarity diagram, this paper first describes the ontology of content metadata, and then calculates the similarity to ensure the accuracy of copyright content description and the comprehensiveness of mining. The node relational connection module records the nature of message timestamp, message frequency and message type to complete the quantification of message weight. The process of message filtering and filtering greatly simplifies P2P network data and optimizes the implementation of community mining algorithm in advance. The third step is to propose an effective ANMGN algorithm for tort community mining on the basis of studying the classical community mining algorithm Girvan-Newman (GN) algorithm, and introduce the community structure gain to shorten the algorithm time. At present, one of the supervision of copyright content is to deal with it after the event, that is, after the discovery of pirated infringing content, the dissemination of infringing content is blocked by technology such as preventing download, file removal and so on. The other is content format encryption technology, which can only be downloaded by users licensed by copyright. This technology is not open source, it is only suitable for specific software, and it is not universal. The method based on infringing community mining proposed in this paper realizes the detection and supervision of the dissemination direction and distribution of copyright content, and achieves the regional early warning to crack down on the infringing community in time. In order to solve the problem of the dissemination of pirated infringing content in P2P network, this paper analyzes the threat degree of the community on the basis of the discovery of the infringing community. This paper puts forward two kinds of countermeasures for the dissemination of copyright content in infringing community: the strategy of breaking point and the strategy of breaking edge. When the copyright content propagates, the node that passes through the most times, we call it the bridge point, the breakpoint reaction strategy is to find the community bridge point, delete these bridge points when necessary, in order to prevent the copyright content from spreading outward. The connecting edge between communities, which we call key edge, breaks the edge strategy by finding the key edge in the network and producing these edges if necessary, which destroys the dissemination of copyright content.
【學(xué)位授予單位】：山東大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TP393.02

【參考文獻】

相關(guān)期刊論文 前5條

1 高琦,陳華鈞;互聯(lián)網(wǎng)Ontology語言和推理的比較和分析[J];計算機應(yīng)用與軟件;2004年10期

2 竇文,王懷民,賈焰,鄒鵬;構(gòu)造基于推薦的Peer-to-Peer環(huán)境下的Trust模型[J];軟件學(xué)報;2004年04期

3 朱曉姝;;一種語義對等網(wǎng)模型[J];網(wǎng)絡(luò)安全技術(shù)與應(yīng)用;2006年03期

4 李秀蓮;P2P軟件使用用戶版權(quán)侵權(quán)問題的解決對策[J];現(xiàn)代情報;2005年08期

5 徐德智;汪智勇;王斌;;當(dāng)前主要本體推理工具的比較分析與研究[J];現(xiàn)代圖書情報技術(shù);2006年12期

，

本文編號：2492370