【正文】 切關(guān)心的熱門話題或反映了某些社會心理的觀點(diǎn)與看法，其較高層次是‘思潮’，基本層次是‘情緒’。 微博 成為網(wǎng)絡(luò)輿情的一大組成部分。美國 埃文 利用 空間向量模型 完成 對微博內(nèi)容的 結(jié)構(gòu) 數(shù)據(jù)化 ， 利用 Kmeans 算法 完成 對微博消息 的 聚類分析， 找到所 要分析的某類微博內(nèi)容，進(jìn)而在這類微博中 找出微博消息意見領(lǐng)袖，提出微博 意見 領(lǐng)袖 影響力 評估算法， WeiboRank 算法， 并 結(jié)合 算法完成了微博消息預(yù)警模 塊的實(shí)現(xiàn) ， 初步實(shí)現(xiàn)了微博輿情管理平臺的數(shù)據(jù) 預(yù)警 分析 功能 。北京交通大學(xué)畢業(yè)設(shè)計（論文） 畢業(yè)設(shè)計（論文） 中文題目： 微博輿情管理平臺 數(shù)據(jù)分析系統(tǒng)的設(shè)計與實(shí)現(xiàn) 英文題目 : MicroBlog Public Opinion Management Platform: The Design and Implementation of the Data Analysis System 1 中文摘要 隨著網(wǎng)絡(luò)技術(shù)應(yīng)用的普及和發(fā)展，輿情的傳播方式和傳播速度都發(fā)生了根本性變化 , 網(wǎng)絡(luò)輿情對人類的社會狀態(tài)產(chǎn)生了全方位的影響，微博 輿情則是網(wǎng)絡(luò)輿情的重要組成部分， 它 的特點(diǎn)有： 直接性，突發(fā)性，偏差性，豐富性和互動性。 關(guān)鍵詞： 微博輿情 文本聚類 趨勢分析 北京交通大學(xué)畢業(yè)設(shè)計（論文） Abstract Along with the universal application and rapid development of work technology, the approaches that the mediated public sentiment spread have been fundamentally changed. The mediated public sentiment has exerted huge influence on the way that the society operates. As the one of the most significant parts of the mediated public sentiment, the public sentiment which is produced and spread by the microblog has several important characters, such as directness, immediacy, deviation, variability, interactivity. Taking the microblog messages as our investigating subject, this paper aimed to do research on the characteristics and models of delivering messages between microblog users, Through the analysis of the capture data found unidirectional, microblog munication convenience, back on the face and other characteristics, and raised an effective algorithm to sort these kinds of messages. Using the spatial vector model, the Kmeans algorithm did cluster analysis on microblog messages, and found out the opinion leaders among tremendous messages. Then, an influential estimation algorithm of the microblog opinion leaders was raised， WeiboRank algorithm. Together with the estimation algorithm, we also achieved the early warning part and some basic data warning analysis functions on the whole microblogmediated public sentiment platform. Key words： microblogmediated public sentiment, text clustering, trend analysis 北京交通大學(xué)畢業(yè)設(shè)計（論文） 目 錄 一、 概述 ..................................................................................................... 1 課題背景與研究意義 ........................................................................ 1 課題背景 .................................................................................. 1 研究現(xiàn)狀 .................................................................................. 3 研究意義 .................................................................................. 3 論文結(jié)構(gòu) ............................................................................................. 4 二、微博消息傳播模型 ................................................................................... 4 微博消息傳播的特點(diǎn) ......................................................................... 4 微博用戶狀態(tài) ..................................................................................... 6 微博意見領(lǐng)袖 ..................................................................................... 7 微博傳播模型 ..................................................................................... 9 三、微博輿情管理平臺的設(shè)計與實(shí)現(xiàn) ......................................................... 12 微博輿情管理平臺的總體流程 ....................................................... 12 數(shù)據(jù)分析系統(tǒng)設(shè)計流程 ................................................................... 13 四、微博輿情管理平臺的實(shí)現(xiàn) ..................................................................... 14 樣本選取與數(shù)據(jù)來源 ....................................................................... 14 微博數(shù)據(jù)轉(zhuǎn)化 ................................................................................... 15 微博文本聚類 ................................................................................... 17 文本聚類定義 ........................................................................ 17 機(jī)器學(xué)習(xí) ................................................................................ 18 算法 .......................................................................... 19 微博意見領(lǐng)袖重要性評估 ............................................................... 21 PageRank 算法 ....................................................................... 21 WeiboRank 算法 ..................................................................... 22 算法對比 ............................................................................... 23 微博輿情預(yù)警模塊 ........................................................................... 25 微博輿情預(yù)警 ........................................................................ 25 北京交通大學(xué)畢業(yè)設(shè)計（論文） 趨勢分析模 塊 ........................................................................ 26 趨勢分析結(jié)果比較 ........................................................................... 29 五、結(jié)論與展望 ............................................................................................. 31 系統(tǒng)不足 ........................................................................................... 31 未來展望 ........................................................................................... 32 改進(jìn)預(yù)期 ................................................................................ 32 新增功能 ................................................................................ 32 結(jié)束語 ............................................................................................... 33 參考文獻(xiàn) ......................................................................................................... 34 附錄Ⅰ： 翻譯原文 ..................................................................................... 35 Cluster Analysis:Basic Concepts and Algorithms ........................................... 35 1Overview ...................................................................................................