本文關鍵詞：社交網(wǎng)絡特征計算與關鍵節(jié)點識別的實驗研究　出處：《吉林大學》2015年碩士論文　論文類型：學位論文

  更多相關文章： 社交網(wǎng)絡 NetworkX 復雜網(wǎng)絡理論 PageRank算法

【摘要】：隨著web2.0的發(fā)展,社交網(wǎng)絡也得到飛速的壯大,各類社交網(wǎng)站和服務的出現(xiàn),不僅極大地豐富了人們的生活,同時也對人類的社交行為和生活方式產(chǎn)生了深刻改變。對社交網(wǎng)絡相關領域的研究是當前網(wǎng)絡研究的熱點。因為社交網(wǎng)絡的獨特地位和作用,已經(jīng)深刻影響了人們社會生活的方方面面，同時也關系到輿論導向、社會安全,甚至國家安全.因此,對社交網(wǎng)絡的研究具有重大意義。 本文針對社交網(wǎng)絡,基于復雜網(wǎng)絡理論和改進的ADWP (Activity degree weighted Pagerank)算法,使用圖論與復雜網(wǎng)絡建模工具Networkx,對社交網(wǎng)絡的特征和關鍵節(jié)點識別進行了實驗分析和研究。本文的主要工作包括： 1.基于復雜網(wǎng)絡理論,針對社交網(wǎng)絡的拓撲結構,以騰訊微博為代表,計算了相關的無標度和小世界特征,在Networkx平臺,驗證了這兩個特征,并進行了實驗分析。 2.基于變化的拓撲結構思想,采用刪除法和收縮法對節(jié)點的重要性進行評估,提出了改進的ADWP關鍵節(jié)點識別算法。本文首先詳細介紹和分析了基于PageRank的關鍵節(jié)點識別算法,然后,在此基礎上,引入社交網(wǎng)絡中用戶“活躍度”這個重要因素,在權重分配上進行了改進。即在基于PageRank的依據(jù)鏈接進行分配的基礎上增加了用戶的“活躍度”作為權重分配指標,更好的刻畫出社交網(wǎng)絡的特征,改進和完善了原有的PageRank關鍵節(jié)點識別算法。最后,在Networkx環(huán)境下以騰訊微博的轉發(fā)網(wǎng)絡為例,驗證了算法的基本思想和識別關鍵節(jié)點的效果。 綜上所述,本文基于Networkx環(huán)境,從社交網(wǎng)絡的復雜網(wǎng)絡特征計算和關鍵節(jié)點識別發(fā)現(xiàn)兩個方面進行了實驗分析和研究。本文的研究工作具有一定前沿性,對同類工作,也具有一定的理論參考價值。
[Abstract]:With the development of web2.0, social network has been growing rapidly. The emergence of various social networking sites and services has not only greatly enriched people's lives. At the same time, the social behavior and lifestyle of human beings have been profoundly changed. The research on social network is the focus of the current network research, because of the unique status and role of social networks. It has deeply affected all aspects of people's social life, but also related to the guidance of public opinion, social security, and even national security. Therefore, the study of social networks is of great significance. This paper aims at social network, based on complex network theory and improved ADWP activity degree weighted algorithm. With the help of graph theory and complex network modeling tool Networkx, the characteristics and key node identification of social networks are analyzed and studied experimentally. The main work of this paper is as follows: 1. Based on the theory of complex network, aiming at the topology of social network, taking Tencent Weibo as the representative, the scale-free and small-world features are calculated, and the two features are verified on the Networkx platform. Experimental analysis was also carried out. 2. Based on the idea of changing topology, the importance of nodes is evaluated by deleting and shrinking methods. An improved ADWP key node recognition algorithm is proposed. Firstly, this paper introduces and analyzes the key node recognition algorithm based on PageRank in detail, and then, on this basis. Introduction of the social network user "activity" is an important factor. The weight distribution is improved, that is, the user's "activity" is added as the index of weight allocation on the basis of link allocation based on PageRank. Better portray the characteristics of social networks, improve and improve the original PageRank key node identification algorithm. Finally, in the Networkx environment, Tencent Weibo forwarding network as an example. The basic idea of the algorithm and the effect of identifying key nodes are verified. To sum up, this article is based on the Networkx environment. The experiment analysis and research are carried out from the two aspects of complex network feature calculation and key node recognition of social network. The research work in this paper has some vanguard, and the same kind of work has been done. Also has certain theoretical reference value.
【學位授予單位】：吉林大學
【學位級別】：碩士
【學位授予年份】：2015
【分類號】：TP393.09

【參考文獻】

相關期刊論文 前5條

1 才華;周春光;王U，

本文編號：1378429