本文選題：數(shù)據(jù)挖掘　切入點：聚類分析　出處：《浙江工業(yè)大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：近兩年來浙江經(jīng)濟形勢未有明顯改觀,銀行資產(chǎn)質(zhì)量持續(xù)下行,不良率高企,嚴(yán)重侵蝕銀行利潤,銀行亟待提升對客戶的風(fēng)險管理水平。當(dāng)前銀行對私客戶的個人貸款業(yè)務(wù)和信用卡業(yè)務(wù)量在全行資產(chǎn)業(yè)務(wù)量中占比近為四成,面臨了較大的信用風(fēng)險、市場風(fēng)險和操作風(fēng)險,機構(gòu)涉及面廣,對整體的資產(chǎn)質(zhì)量影響較大。傳統(tǒng)的風(fēng)險管理辦法基于常規(guī)的思路,且多為事后管理性質(zhì),使用的技術(shù)手段相對落后,對風(fēng)險點的監(jiān)控也屬于零散或者是無規(guī)律,缺乏系統(tǒng)性管理,已不能滿足日益迅猛發(fā)展的銀行業(yè)務(wù)。本文在對數(shù)據(jù)挖掘相關(guān)理論大量文獻研究綜述的基礎(chǔ)上,將對私客戶在上述兩種業(yè)務(wù)中有意無意表達出的行為進行分析,通過數(shù)據(jù)清理和轉(zhuǎn)化,為數(shù)據(jù)挖掘提供規(guī)范化數(shù)據(jù),并使用IBM的SPSS Modeler數(shù)據(jù)挖掘工具,通過聚類和決策樹等建模方法,對客戶行為進行了深度分析和挖掘,試圖找出形成風(fēng)險的影響因素,分析偏離正常值的異常信息,形成有參考價值的模型用于預(yù)測未來的風(fēng)險以及監(jiān)控當(dāng)前的風(fēng)險。最后依據(jù)數(shù)據(jù)的可獲取性,以及對數(shù)據(jù)挖掘形成模型的研究后,提出對對私客戶的行為監(jiān)測指標(biāo),以供對口條線管理部門參考,提高管理層決策水平,為信貸政策提供準(zhǔn)確的數(shù)據(jù)基礎(chǔ),并提供科學(xué)依據(jù),制定業(yè)務(wù)發(fā)展目標(biāo)。
[Abstract]:In the past two years, the economic situation in Zhejiang has not improved significantly. The quality of bank assets has continued to decline and the rate of failure has remained high, which has seriously eroded bank profits. Banks urgently need to improve the level of risk management for customers. At present, the amount of personal loans and credit card business of banks to private customers accounts for nearly 40% of the bank's total assets business, and they are faced with greater credit risk, market risk and operational risk. The traditional risk management methods are based on conventional thinking, and most of them are ex post management, and the technical means used are relatively backward. The monitoring of risk points is also scattered or irregular, lack of systematic management, can not meet the rapid development of banking business. This paper on the basis of a large number of data mining related theory literature review, This paper will analyze the behavior expressed by private customers in the above two kinds of business, provide standardized data for data mining through data cleaning and transformation, and use the SPSS Modeler data mining tool of IBM. Through clustering and decision tree modeling methods, this paper makes a deep analysis and mining of customer behavior, tries to find out the influencing factors of risk, and analyzes the abnormal information that deviates from the normal value. A model with reference value is formed to predict future risks and monitor current risks. Finally, according to the availability of data and the research of data mining model, the behavior monitoring index for private customers is proposed. In order to provide reference for the corresponding line management department, improve the management decision-making level, provide accurate data basis for credit policy, and provide scientific basis to formulate business development goals.
【學(xué)位授予單位】：浙江工業(yè)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TP311.13;F832.2

【參考文獻】

相關(guān)碩士學(xué)位論文 前1條

1 黃偉豪;數(shù)據(jù)挖掘技術(shù)在網(wǎng)上銀行促銷活動中的應(yīng)用[D];華南理工大學(xué);2012年

，

本文編號：1570036