【正文】 是大眾密切關心的熱門話題或反映了某些社會心理的觀點與看法，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 對微博內容的 結構 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容，進而在這類微博中 找出微博消息意見領袖，提出微博 意見 領袖 影響力 評估算法， WeiboRank 算法， 并 結合 算法完成了微博消息預警模 塊的實現(xiàn) ， 初步實現(xiàn)了微博輿情管理平臺的數(shù)據(jù) 預警 分析 功能 。北京交通大學畢業(yè)設計（論文） 畢業(yè)設計（論文） 中文題目： 微博輿情管理平臺 數(shù)據(jù)分析系統(tǒng)的設計與實現(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)絡技術應用的普及和發(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)絡輿情的重要組成部分， 它 的特點有： 直接性，突發(fā)性，偏差性，豐富性和互動性。 關鍵詞： 微博輿情 文本聚類 趨勢分析 北京交通大學畢業(yè)設計（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 北京交通大學畢業(yè)設計（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 論文結構 ............................................................................................. 4 二、微博消息傳播模型 ................................................................................... 4 微博消息傳播的特點 ......................................................................... 4 微博用戶狀態(tài) ..................................................................................... 6 微博意見領袖 ..................................................................................... 7 微博傳播模型 ..................................................................................... 9 三、微博輿情管理平臺的設計與實現(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)設計流程 ................................................................... 13 四、微博輿情管理平臺的實現(xiàn) ..................................................................... 14 樣本選取與數(shù)據(jù)來源 ....................................................................... 14 微博數(shù)據(jù)轉化 ................................................................................... 15 微博文本聚類 ................................................................................... 17 文本聚類定義 ........................................................................ 17 機器學習 ................................................................................ 18 算法 .......................................................................... 19 微博意見領袖重要性評估 ............................................................... 21 PageRank 算法 ....................................................................... 21 WeiboRank 算法 ..................................................................... 22 算法對比 ............................................................................... 23 微博輿情預警模塊 ........................................................................... 25 微博輿情預警 ........................................................................ 25 北京交通大學畢業(yè)設計（論文） 趨勢分析模 塊 ........................................................................ 26 趨勢分析結果比較 ........................................................................... 29 五、結論與展望 ............................................................................................. 31 系統(tǒng)不足 ........................................................................................... 31 未來展望 ........................................................................................... 32 改進預期 ................................................................................ 32 新增功能 ................................................................................ 32 結束語 ............................................................................................... 33 參考文獻 ......................................................................................................... 34 附錄Ⅰ： 翻譯原文 ..................................................................................... 35 Cluster Analysis:Basic Concepts and Algorithms ........................................... 35 1Overview ...............................................................................................