本文選題：新浪微博 + 用戶關(guān)系��； 參考：《西安電子科技大學(xué)》2014年碩士論文

【摘要】：隨著信息技術(shù)產(chǎn)業(yè)的快速發(fā)展,寬帶網(wǎng)絡(luò)和高速移動網(wǎng)絡(luò)的速度大幅提升,各種新型網(wǎng)絡(luò)接入終端設(shè)備也變得越來越普及,極大地提高了人們通過網(wǎng)絡(luò)進(jìn)行溝通交流的速度和頻率。當(dāng)下,微博已成為網(wǎng)絡(luò)思想交流的重要平臺、網(wǎng)絡(luò)輿情的高度聚集地,是人們交流思想、政府洞察民意的重要窗口。高用戶活躍度和大量博文背后隱藏著重大的數(shù)據(jù)價(jià)值,通過代理數(shù)據(jù),準(zhǔn)確理解微博用戶的交互行為、挖掘用戶關(guān)系等數(shù)據(jù)中所蘊(yùn)含的信息和影響具有重要的意義。本文基于微博平臺對其所包含的用戶關(guān)系進(jìn)行分析和研究,主要研究內(nèi)容和研究成果如下:1.通過對目前微博用戶關(guān)系研究現(xiàn)狀進(jìn)行整理和分析,選定人物關(guān)系強(qiáng)度作為研究對象,從新浪微博的用戶關(guān)注和互動關(guān)系出發(fā),設(shè)計(jì)了人物關(guān)系強(qiáng)度預(yù)測模型。虛擬社交網(wǎng)絡(luò)是現(xiàn)實(shí)社會社交網(wǎng)絡(luò)的映射,其人物間的關(guān)系強(qiáng)度也具有相似的影響因素。因此,此模型通過對現(xiàn)實(shí)世界中人與人之間關(guān)系強(qiáng)度影響因素進(jìn)行全面考慮,并將其遷移應(yīng)用至新浪微博平臺,抽取微博信息中反映用戶間關(guān)系強(qiáng)度的可用信息,使用相似度、標(biāo)準(zhǔn)差等數(shù)學(xué)算法或概念將反映用戶關(guān)系強(qiáng)度的可用信息轉(zhuǎn)變?yōu)榫唧w的數(shù)值信息,最終通過線性模型將多個(gè)影響因素綜合從而進(jìn)行定量化分析。2.實(shí)現(xiàn)了基于微博的人物關(guān)系強(qiáng)度預(yù)測系統(tǒng)。本系統(tǒng)中最重要的數(shù)據(jù)采集部分使用基于模擬登錄和基于新浪微博API兩種方式采集數(shù)據(jù),實(shí)現(xiàn)了微博用戶關(guān)系信息、用戶個(gè)人資料信息和用戶微博內(nèi)容的自動提取。同時(shí),綜合使用上述兩種數(shù)據(jù)采集方法,不僅避免了直接使用API的數(shù)據(jù)獲取限制和用戶未登錄所造成的網(wǎng)頁內(nèi)容獲取數(shù)據(jù)的不完整的問題,而且降低了大量數(shù)據(jù)分析和提取的工作量。3.人物關(guān)系強(qiáng)度預(yù)測系統(tǒng)根據(jù)設(shè)計(jì)的人物關(guān)系強(qiáng)度預(yù)測模型,對系統(tǒng)中數(shù)據(jù)采集部分獲取的數(shù)據(jù)進(jìn)行整理、分析和計(jì)算,預(yù)測人物關(guān)系強(qiáng)度,同時(shí)通過圖形化界面展示人物關(guān)系強(qiáng)度。最后將某微博用戶的預(yù)測模型結(jié)果與新浪微博人脈關(guān)系示例進(jìn)行比較,證明了所設(shè)計(jì)的人物關(guān)系強(qiáng)度模型的有效性。本文所研究的人物關(guān)系強(qiáng)度能夠?qū)ξ⒉┲械挠脩暨M(jìn)行更準(zhǔn)確的親疏關(guān)系劃分,基于本文的研究能夠支持進(jìn)一步使用社團(tuán)分區(qū)算法進(jìn)行更高準(zhǔn)確度的好友推薦;支持輿情的精準(zhǔn)發(fā)現(xiàn),在輿情預(yù)警機(jī)制的使用中提高輿情預(yù)警的準(zhǔn)確度;支持向用戶推薦不同好友的隱私保護(hù)策略,幫助識別用戶好友,同時(shí)保護(hù)用戶隱私。
[Abstract]:With the rapid development of the information technology industry, the speed of broadband network and high speed mobile network has been greatly improved. All kinds of new network access terminal equipment have become more and more popular, which has greatly improved the speed and frequency of communication and communication through the network. At present, micro-blog has become an important platform for network thought communication, network public opinion The highly aggregated area is an important window for people to exchange ideas and the government's insight into the public opinion. High user activity and a large number of blog posts are hidden with significant data value. Through proxy data, it is important to understand the interactive behavior of micro-blog users and to excavate the information and influence in the data of user relations. The main research content and research results are as follows: 1. by sorting and analyzing the current research status of micro-blog user relations, the relationship strength of the characters is selected as the research object, and the relationship intensity of sina micro-blog is designed to predict the intensity of the relationship. Model. Virtual social network is the mapping of social social network in real society, and the relationship intensity of human and object has similar influence factors. Therefore, this model is fully considered by the influence factors of the relationship intensity between people in the real world, and applies its migration to the Sina micro-blog platform and extracts micro-blog information to reflect the users. Available information of relationship intensity, using mathematical algorithms or concepts such as similarity, standard deviation and other mathematical algorithms to convert the available information of user relationship strength into specific numerical information. Finally, a number of factors are synthesized by linear model and then quantificationally analyzed by.2.. In this system, a micro-blog based relationship strength prediction system is implemented. The most important data acquisition part uses two kinds of data acquisition based on analog logon and Sina micro-blog API. It realizes the micro-blog user relationship information, the user's personal information and the automatic extraction of the content of the user's micro-blog. At the same time, the comprehensive use of the above two data acquisition methods not only avoids the direct use of data acquisition limits for the use of the data acquisition methods. And the incompleteness of data obtained from the content of web pages caused by users not logged in, and reducing a large amount of data analysis and extraction of the workload.3. character relationship intensity prediction system based on the predicted model of the relationship strength of personage, the data collected in the data collection part of the system are collated, analyzed and calculated, and the figure is predicted. In the end, a micro-blog user's prediction model is compared with the Sina micro-blog relationship example, and the validity of the model is proved. The relationship strength studied in this paper can be more accurate and close to the users in the micro-blog. Relationship division, based on this study, it can support the further use of community partition algorithm for better friend recommendation, support the accurate discovery of public opinion, improve the accuracy of public opinion early warning in the use of public opinion early warning mechanism, support the privacy protection strategies of different friends to users, help identify the user friends, and protect the users. Protect the user's privacy.
【學(xué)位授予單位】：西安電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TP393.092

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 陳天;劉文浩;;相似度算法分析與比較研究[J];現(xiàn)代計(jì)算機(jī)(專業(yè)版);2012年18期

，

本文編號：2100355