發(fā)布時間：2018-04-09 14:00

  本文選題：群體性突發(fā)事件　切入點：預警　出處：《北京郵電大學》2012年碩士論文

【摘要】：隨著我國改革開放的深化和市場經濟的發(fā)展,很多新問題、新矛盾以不同的形式表現(xiàn)出來。特別是近年來我國群體性突發(fā)事件頻頻發(fā)生,造成的后果和影響也越來越嚴重,同時隨著互聯(lián)網的飛速發(fā)展,使得群體性事件的爆發(fā)更加突然,不斷威脅到我國社會和諧,引起了社會各界的密切關注和深入思考。對于很多傳統(tǒng)的群體性事件,可利用數(shù)據(jù)挖掘相關技術進行監(jiān)控提取,但是近期出現(xiàn)的這樣一類群體性事件,呈現(xiàn)出概率小、突發(fā)性、隱匿性等特點,其在產生之初,被掩蓋在大量的平凡數(shù)據(jù)中,如用傳統(tǒng)算法進行抽取,則會在數(shù)據(jù)處理初期被當成噪聲數(shù)據(jù)濾除,從而沒有引起足夠的重視。如果忽視這類群體性突發(fā)事件,將會產生非常惡劣的后果。 針對以上問題,本課題通過對這類群體性突發(fā)事件的研究,力圖利用技術手段解決其發(fā)現(xiàn)預警不足的問題,通過整體建模和軟件實現(xiàn),期望能夠很好的實現(xiàn)群體性突發(fā)事件的預警。同時,結合實際項目需求,將這個模型應用到基于某單位投訴文本的群體性突發(fā)事件分析預警系統(tǒng)中,用于預警某單位業(yè)務中出現(xiàn)的新問題和群體性事件。完成系統(tǒng)設計與實現(xiàn)的同時對下一步的深入研究也提出了一些建議。 本文重點闡述了以下幾個方面的工作： 1.通過分析得出這類群體性突發(fā)事件的定義、相關特性,并在此基礎上提出了相應的數(shù)據(jù)模型,對群體性突發(fā)事件進行描述。 2.加強了對數(shù)據(jù)來源的分析,尤其是加入了網絡數(shù)據(jù)作為數(shù)據(jù)來源的一部分,同時本文提出了一種基于滑動窗口模型的挖掘算法以去除噪聲數(shù)據(jù)。 3.提出一種利用遺傳算法優(yōu)化的K-means聚類方法,以更好的實現(xiàn)聚類分析,提高了結果的準確性、可靠性。 4.將此分析預警模型應用到電信領域,結合與某單位合作的項目,對其相關投訴數(shù)據(jù)進行處理,實現(xiàn)對電信投訴的群體性突發(fā)事件預警。
[Abstract]:With the deepening of China's reform and opening up and the market economy, many new problems and new contradictions in different forms. Especially in recent years, the sudden mass incidents happen frequently in China, the consequences and impact is more and more serious, at the same time, with the rapid development of the Internet, the outbreak of mass incidents more suddenly and continue to threaten the harmony of our society, attracted close attention and deep thinking of the community. For many traditional group events, can use data mining technology to monitor the extraction, but the recent emergence of such group events, showing a small probability, sudden, characteristics of occult, the the beginning, hidden in the ordinary large quantities of data, such as the use of traditional algorithms for extraction, data processing will be in the early stage as noise data filtering, and if not attracted enough attention. The neglect of such mass emergencies will have very bad consequences.
To solve the above problems, through the research of this kind of group incidents, trying to use technical means to solve the warning problems, through the realization of integrated modeling and software, hoping to achieve mass emergency warning as well. At the same time, combined with the actual needs of the project, the model is applied to a group the text of the emergency Complaints Unit Based on the analysis of the early warning system for the event, new problems and group appear warning a unit in the business. At the same time to complete the system design and implementation of the next step of the research also puts forward some suggestions.
This article focuses on the following aspects of the work:
1., through the analysis, we get the definition and characteristics of such group emergencies, and on this basis, we propose corresponding data models to describe group emergencies.
2., we strengthened the analysis of data sources, especially joined the network data as a part of data sources. Meanwhile, we proposed a mining algorithm based on sliding window model to remove noise data.
3. a K-means clustering method optimized by genetic algorithm is proposed to improve the accuracy and reliability of the cluster analysis.
4., apply the analysis and early warning model to the field of telecommunications, and deal with the related complaints data combined with a cooperation project with a unit, so as to realize the early warning of mass incidents of Telecom complaints.

【學位授予單位】：北京郵電大學
【學位級別】：碩士
【學位授予年份】：2012
【分類號】：TP311.13

【引證文獻】

相關碩士學位論文 前1條

1 段曉燕;投訴信息自動分類與推送系統(tǒng)的研究與設計[D];北京郵電大學;2013年

，

本文編號：1726728