本文關鍵詞： 信息記錄平臺 B/S架構 J2EE框架 聚類方法 劃分方法　出處：《吉林大學》2017年碩士論文　論文類型：學位論文

【摘要】：生活在21世紀的我們,已經(jīng)處在一個高度發(fā)達的信息科學技術時代,信息科學技術的發(fā)展快速無比,尤其在經(jīng)濟、政治、文化、軍事等方面的應用更是與日俱增,信息技術在使得人類社會文明取得了很大的進步與發(fā)展的同時,在互聯(lián)網(wǎng)金融行業(yè)的應用也逐漸廣泛。為了高效地管理人們信用卡的消費信息,提高信息處理效率,信用卡信息記錄平臺已逐漸地應用到企業(yè)對信用卡信息的管理當中。但在信息技術廣泛應用的同時,數(shù)據(jù)庫中存儲的信息量也快速增長,使得人們高效提取有效信息的難度逐步增加,由此導致對數(shù)據(jù)挖掘技術應用的迫切需求,將數(shù)據(jù)挖掘算法應用在信息量較大的平臺上也是非常有必要的。近些年來,數(shù)據(jù)挖掘已經(jīng)逐漸被應用到各個行業(yè)領域當中,例如互聯(lián)網(wǎng)金融、車載網(wǎng)、電子地圖導航應用等各個領域。作為當前的前沿科研課題與很有發(fā)展前景的技術,數(shù)據(jù)挖掘不僅涉及到數(shù)據(jù)庫技術、機器學習等多種技術,還與統(tǒng)計學、概率理論等多種基礎理論緊密相關。聚類分析算法是數(shù)據(jù)挖掘方法之一,近年來人們在這方面的研究探索取得了很大的進步,不同的聚類分析方法也紛紛出現(xiàn),提出了各種基于不同思想的聚類方法,并加以驗證,給出實現(xiàn),最后應用在合適的場景之中。在涉及到人工智能科學技術的各個領域當中,這些方法都有所涉及,在相應的領域當中也取得了眾多成果。然而,當下的互聯(lián)網(wǎng)金融行業(yè)之中,有一些問題需要解決,更需合適的聚類方法應用于其中。本文工作主要是構建信用卡信息記錄平臺,選擇、改進聚類方法,最后將其應用于平臺。具體內容主要包括:(1)構建了信用卡信息記錄平臺。基于B/S架構與Java平臺,針對目標需求,進行開發(fā),設計了對應的功能模塊,分別為安全管理、信息采集、信息管理、統(tǒng)計報表、信息聚類和平臺管理等模塊,支持根據(jù)需求開發(fā)新的功能模塊。前端使用HTML、CSS、Java Script設計頁面,后端使用Java語言編寫,并采用Struts2、Hibernate、Spring等框架的相關技術,數(shù)據(jù)庫選擇在互聯(lián)網(wǎng)上應用比較廣泛并且開源的My SQL。(2)深入研究了基于劃分的聚類算法。劃分的方法主要有基于k平均的和k中心的方法,在設計的信息平臺上,信用分是作為衡量消費者信用的主要方式,從上限100遞減,每個分數(shù)段作為一個信用等級劃分,到下限60為止。雖然k中心算法所劃分的數(shù)據(jù)結果明顯,由于其隨機選擇初始中心點,聚類結果具有不可控性,每一組聚類結果分布與每一個信用等級范圍不是很接近。而本文通過對傳統(tǒng)的k中心算法進行改進,選取信息記錄條數(shù)最多的k組數(shù)據(jù)作為中心點,同時通過對迭代次數(shù)的限制,然后進行聚類劃分,可以提高接近程度來提高管理消費者信用等級的效率,從而高效地觀察不同信用等級消費者的行為習慣,制定相應的信用等價折扣策略,提高商業(yè)利益。本文最后對文中的工作進行了總結,對算法的優(yōu)缺點給出了分析,并對未來的工作做出了展望。
[Abstract]:We have been living in the twenty-first Century, in a highly developed information technology era, the rapid development of information science and technology incomparable, especially in the economic, political, cultural, military and other aspects of the application of information technology is in progress and grow with each passing day, the development of the civilization of human society made great and application in Internet banking the industry is gradually widely. In order to efficiently manage the credit card consumer information, improve the efficiency of information processing, credit card information recording platform has been gradually applied to the enterprise credit card information management. But in the extensive application of information technology at the same time, the amount of information stored in the database is growing rapidly, which makes people efficient extraction the effective information is more difficult, which led to an urgent need for application of data mining, data mining algorithm application platform in large amount of information Is also very necessary. In recent years, data mining has been applied to various fields, such as Internet banking, mobile network, the electronic map navigation application. As a frontier research topic of current and promising technology, data mining is not only related to the database technology, machine learning and other technologies also, with statistics, probability theory and other basic theory are closely related. Clustering algorithm is one of the methods of data mining in recent years, people explore the research in this area has made great progress, with the method of clustering analysis have also appeared, put forward a variety of different clustering methods based on the idea, and verified, and presents the finally, the appropriate application in the scene. In artificial intelligence involves various fields of science and technology, these methods have been involved in the field, when Also made many achievements. However, in the current Internet financial industry, some problems need to be solved, more suitable to the application of clustering method. The main work of this paper is to construct the credit card information recording platform selection, improved clustering method, and finally applied to the platform. The specific contents include: (1) the construction of the credit card information recording platform. The architecture of B/S and Java based platform, development target, demand, function module design of the corresponding, respectively for the safety management, information collection, information management, statistical statements, information clustering and platform management module, support the development of new functional modules using HTML front-end according to requirements. CSS, Java, Script page design, is the use of Java language, and the use of Struts2, Hibernate, Spring and other related technical framework, database is widely used on the Internet and open source My S QL. (2) studied clustering algorithm based on the division. The division methods are mainly K and K based on the average center, in the design of the information platform, credit is as the main measure of consumer credit, decrease from the upper limit of 100, each fraction as a credit rating, to the lower limit 60 so far. Although the k center algorithm divided data the results are obvious, because of its random selection of initial centers, the clustering result is not controllable, the distribution of each group clustering and each credit rating range is not very close. And through the center of the traditional K algorithm is improved, the number of records selected information K group most data as the center point, at the same time through the iteration limit, then clustering can improve the proximity to improve the efficiency of the management of consumer credit rating, to effectively observe different credit rating Finally, we summarize the work in this paper, analyze the advantages and disadvantages of the algorithm, and make prospects for future work.

【學位授予單位】：吉林大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP311.13

【參考文獻】

相關期刊論文 前9條

1 阮菊紅;丁衛(wèi)撐;李勇;;論知識管理對信息資源管理的深化與拓展[J];蘭臺世界;2009年04期

2 劉華星;楊庚;;HTML5——下一代Web開發(fā)標準研究[J];計算機技術與發(fā)展;2011年08期

3 高兵;鄒啟杰;;基于網(wǎng)格的聚類方法研究[J];軟件工程師;2010年Z1期

4 趙恩來;郝文寧;趙水寧;韓憲勇;;改進的基于密度方法的態(tài)勢聚類顯示算法[J];計算機工程;2010年18期

5 劉艷春;洪曉慧;;Struts2框架核心配置文件的研究與應用[J];計算機技術與發(fā)展;2013年02期

6 林土方;洪凱星;郭才福;潘成峰;王澤波;陳祥獻;黃海;;基于B/S架構的變壓器在線狀態(tài)監(jiān)測系統(tǒng)實現(xiàn)[J];電子測量與儀器學報;2013年08期

7 陳正舉;;基于HIBERNATE的數(shù)據(jù)庫訪問優(yōu)化[J];計算機應用與軟件;2012年07期

8 王先文;陳鋒;程智;杜耀華;暴洪濤;吳太虎;;基于偏斜t混合模型的流式數(shù)據(jù)自動聚類方法研究[J];電子學報;2014年12期

9 查修齊;吳榮泉;高元鈞;;C/S到B/S模式轉換的技術研究[J];計算機工程;2014年01期

相關碩士學位論文 前10條

1 陳曉;互聯(lián)網(wǎng)金融對商業(yè)銀行盈利的影響和實證分析[D];北京外國語大學;2016年

2 杜巖;吉林省村鎮(zhèn)信息管理平臺的設計與實現(xiàn)[D];吉林大學;2015年

3 劉楠楠;中國互聯(lián)網(wǎng)金融風險分析及監(jiān)管研究[D];山東大學;2014年

4 袁雨田;JavaScript中基于WebCL的并行AES和FFT的設計與實現(xiàn)[D];哈爾濱工業(yè)大學;2014年

5 胡智超;騰訊Q~+應用商店前端框架設計與實現(xiàn)及性能優(yōu)化[D];哈爾濱工業(yè)大學;2013年

6 蘇桁;一個基于WCF和jqGrid的Web應用系統(tǒng)的設計與實現(xiàn)[D];華南理工大學;2012年

7 丁楠;基于HTML5和CSS3的網(wǎng)易WAP站觸屏版移動產(chǎn)品研究[D];華中科技大學;2012年

8 張芳;數(shù)據(jù)挖掘中的聚類分析技術研究[D];國防科學技術大學;2011年

9 何晏成;基于近鄰傳播和凝聚層次的文本聚類方法[D];哈爾濱工業(yè)大學;2010年

10 江圣祺;數(shù)據(jù)挖掘技術在平安銀行信用卡風險管理系統(tǒng)中的應用[D];上海交通大學;2010年

，

本文編號：1466845