本文選題：微博話題 + 熱點預(yù)測��； 參考：《北京郵電大學(xué)》2014年碩士論文

【摘要】：微博作為一種新型的自媒體平臺,以其信息傳播速度快、信息源多、用戶參與度高而得到迅速發(fā)展。同時,微博的這種傳播特性也導(dǎo)致信息的龐大而雜亂,信息的真?zhèn)涡院驮掝}的熱度等成為人們關(guān)注的重點。本文根據(jù)微博分析中的熱門應(yīng)用,研究了微博熱點話題預(yù)測與話題傳播途徑分析方面的相關(guān)技術(shù)。 本文采用話題源頭假設(shè)模型研究新浪微博的熱點話題預(yù)測。該方法將微博用戶群抽象為話題源頭,并根據(jù)熱門話題和非熱門話題由話題源頭產(chǎn)生的假設(shè),將熱點話題預(yù)測問題轉(zhuǎn)化為話題的分類問題。同時,論文實現(xiàn)并改進了一種基于微博話題關(guān)注度的時間序列結(jié)構(gòu)化學(xué)習(xí)的方法,對熱門話題進行預(yù)測。實驗結(jié)果證明該方法能以90%的準確率分出熱門話題,其中77.8%的話題的預(yù)測點提前于其上官方微博熱門話題榜的時間,提前的平均時間為1.54小時。 另外,本文根據(jù)單條微博與微博話題的轉(zhuǎn)發(fā)評論信息,重構(gòu)了話題傳播途徑,并將其可視化呈現(xiàn)。在此基礎(chǔ)上,本文亦對話題傳播途徑中參與用戶的影響力做了分析,實現(xiàn)了一種話題傳播中關(guān)鍵用戶的計算方法,可以有效的進行微博話題溯源分析。
[Abstract]:As a new type of self-media platform, Weibo has been developed rapidly because of its high speed of information dissemination, more information sources and high user participation. At the same time, the spread of Weibo also leads to the huge and messy information, the authenticity of information and the heat of topics become the focus of attention. According to the popular application of Weibo, this paper studies the techniques of Weibo hot topic prediction and topic propagation approach analysis. This paper uses the topic source hypothesis model to study the hot topic prediction of Sina Weibo. This method abstracts the Weibo user group as the topic source and transforms the hot topic prediction problem into the topic classification problem according to the hypothesis that hot topic and non-hot topic are generated by topic source. At the same time, this paper implements and improves a structured learning method of time series based on Weibo topic concern to predict hot topics. The experimental results show that the method can distinguish hot topics with 90% accuracy, 77.8% of the topics are predicted earlier than the official Weibo hot topics list, and the average advance time is 1.54 hours. According to the single Weibo and Weibo topic forwarding comments, this paper reconstructs the topic communication approach and visualizes it. On this basis, this paper also analyzes the influence of the users involved in topic communication, and realizes a calculation method of key users in topic communication, which can effectively analyze the origin of Weibo topics.
【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】：2014
【分類號】：TP393.092

【參考文獻】

相關(guān)期刊論文 前4條

1 高輝;王沙沙;傅彥;;Web輿情的長期趨勢預(yù)測方法[J];電子科技大學(xué)學(xué)報;2011年03期

2 林s，

本文編號：2000573