發(fā)布時(shí)間：2019-04-16 19:41

【摘要】：隨著Web2.0的發(fā)展,在線社交網(wǎng)絡(luò)已逐漸成為互聯(lián)網(wǎng)中最流行的交流平臺。社交網(wǎng)絡(luò)中的用戶可以相互結(jié)識并分享傳播各種信息。社交網(wǎng)絡(luò)中的海量用戶及用戶發(fā)布的信息具有巨大的商業(yè)價(jià)值和研究價(jià)值,這吸引了越來越多的研究者從不同的角度研究社交網(wǎng)絡(luò)。評估社交網(wǎng)絡(luò)節(jié)點(diǎn)影響力是社交網(wǎng)絡(luò)研究的熱點(diǎn)方向之一。通過評估社交網(wǎng)絡(luò)節(jié)點(diǎn)影響力可以發(fā)現(xiàn)社交網(wǎng)絡(luò)中的高影響力人物,對廣告投放、信息管控和用戶行為分析具有重要意義。 本文主要研究社交網(wǎng)絡(luò)節(jié)點(diǎn)影響力評估方法,以更有效地發(fā)現(xiàn)社交網(wǎng)絡(luò)中的高影響力人物。本文在對當(dāng)前評估節(jié)點(diǎn)影響力相關(guān)方法進(jìn)行分析研究的基礎(chǔ)上,提出了一種新的社交網(wǎng)絡(luò)節(jié)點(diǎn)影響力評估方法,并通過實(shí)驗(yàn)對該方法評估節(jié)點(diǎn)影響力的效果進(jìn)行了驗(yàn)證和分析。 針對當(dāng)前有關(guān)社交網(wǎng)絡(luò)中節(jié)點(diǎn)影響力的研究很少關(guān)注影響力在不同主題下的差異以及用戶在不同社區(qū)中擴(kuò)散信息的能力的問題,本文提出了一種新的評估基于主題的節(jié)點(diǎn)影響力的方法。該方法綜合考慮了用戶間傳播信息的可能性和信息被傳播后可能繼續(xù)傳播的范圍,結(jié)合了社交網(wǎng)絡(luò)中用戶間的關(guān)系、用戶的發(fā)布轉(zhuǎn)發(fā)行為和用戶發(fā)布的信息內(nèi)容等因素分析用戶影響力,通過用戶間親密度和社交圈差異度來評估用戶在不同主題下的影響力排名。 本文使用新浪微博中的用戶和微博信息作為實(shí)驗(yàn)數(shù)據(jù)集,對根據(jù)本文方法得到的用戶影響力評估結(jié)果進(jìn)行了分析。實(shí)驗(yàn)表明,本文提出的方法能夠得到不同主題下的用戶影響力排名,不同主題的用戶影響力排名間的相關(guān)度較低,高影響力用戶在多個(gè)社區(qū)中具有更高的信息傳播能力,排名越高的用戶的信息在不同社區(qū)間的傳播越均勻。
[Abstract]:With the development of Web2.0, online social network has gradually become the most popular communication platform in the Internet. Users in social networks can get to know each other and share and spread information. The mass of users in social networks and the information published by users have great business value and research value, which attracts more and more researchers to study social networks from different perspectives. Evaluating the influence of social network nodes is one of the hot topics in the research of social networks. By evaluating the influence of social network nodes, we can find high-impact people in social networks, which is of great significance to advertisement delivery, information control and user behavior analysis. In this paper, we mainly study the evaluation method of social network node influence, in order to find the high-impact person in the social network more effectively. On the basis of the analysis and research of the current methods of evaluating node influence, this paper proposes a new method of evaluating node influence of social network, and validates and analyzes the effect of evaluating node influence by experiments. Current research on node influence in social networks has paid little attention to the differences in influence under different topics and the ability of users to spread information across different communities. In this paper, a new method for evaluating the influence of topic-based nodes is proposed. The method takes into account the possibility of spreading information among users and the extent to which information may continue to spread after it is transmitted, and combines the relationships between users in social networks. The factors such as user's posting and forwarding behavior and the information content published by users are analyzed. The influence ranking of users under different topics is evaluated by the degree of affinity density and social circle difference among users. In this paper, the users and Weibo information in Sina Weibo are used as experimental data sets, and the results of user impact evaluation based on this method are analyzed. Experiments show that the method proposed in this paper can get the ranking of users' influence under different themes, and the correlation between the ranking of users' influence of different topics is low, and the high-impact users have higher information dissemination ability in many communities. The higher the ranking of users, the more evenly they spread across communities.
【學(xué)位授予單位】：北京郵電大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：TP393.09

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 Zhao-yun DING;Yan JIA;Bin ZHOU;Yi HAN;Li HE;Jian-feng ZHANG;;Measuring the spreadability of users in microblogs[J];Journal of Zhejiang University-Science C(Computers & Electronics);2013年09期

，

本文編號：2459042