 |
English
|
正體中文
|
简体中文
|
Items with full text/Total items : 62805/95882 (66%)
Visitors : 3931650
Online Users : 464
|
|
|
Loading...
|
Please use this identifier to cite or link to this item:
https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/31655
|
| Title: | 消費分期市場評等評分模型建置的方向 |
| Other Titles: | The direction of establishing customer's crediting hire-purchase scoring model |
| Authors: | 邱挺晏;Chiu, Ting-yen |
| Contributors: | 淡江大學財務金融學系碩士班
林允永;Lin, Yun-yung |
| Date: | 2005 |
| Issue Date: | 2010-01-11 01:06:09 (UTC+8) |
| Abstract: | 綜觀國際上信用評等的研究與實際運用，已由主觀判斷分析方法和傳統財務比率評分法，轉向為以多變量、依據資本市場理論和利用電腦系統的動態計量分析法為主的發展趨勢，在國內僅有部分的金融業者對於小額信貸與信用卡的申請程序，建立相關的評等評分系統，但在個人消費分期付款上仍無相關公司與單位建立評等評分系統。
在消費分期付款市場中，主要承作對象為一般消費者而一般大眾的價值(Value)難以計算，無法採用市價模式(Mark –to– Market)，故針對個人進行相關的信用評等評分，使用的模型多以違約模式(Default Mode)為主要。而在衡量消費者違約機率的方法中，新興的研究方法中遺傳演算法(genetic algorithms, GAs)、類神經網路(artificial neural network, ANN)、模糊理論 (fuzzy theory)與灰色理論(grey theory)這些方式較少應用於個人風險評估模型中，再加上消費分期付款市場先前尚無發展國內信用評估模型，與消費分期付款市場特性(產品範圍廣但產品深度淺)，若使用上述新興研究方式對此市場中每一個行業設計適用的評等評分制度，將有實務上與時效的限制與困難性。
過去與現階段較常使用的方法中要素評估方式(5C,5P或5S等)容易流於主觀，而多變量信用風險判別模型中logit模型與probit模型的估計結果類似，也就是說logit模型與probit模型所估計的係數方向一致，顯著結果也會相同，只是在實際的數值上有所不同；因此，通常在實證研究上只需要從中選擇其中一種模型即可。
消費分期付款市場若要進行評等評分系統建置，首要任務為資料庫的建立與整合，而分期付款行為與一般放款最大的不同為商品本身的使用與相關後需服務，會影響消費者繳款意願，在進行資料庫建置時，除要納入消費者本身因素外，尚需要考量建立客戶基本資料、信用資料的建檔、交易過程的建立(商品本身&廠商性質)與與客戶聯絡或對保過程等四個類別資料庫的建檔；而資料庫建立、相關系統操作介面與新舊系統間的整合，更是攸關評等評分系統是否可以運作順暢之關鍵。
The credit rating measurement is changing from subjective judgment analysis and traditional financial rate scoring method into multi-variables, CAPM based with computerized system dynamic quantities approach. In Taiwan, there are only a small number of financials apply the credit scoring system on personal credit loan and credit card. There is no company to create the consumer installment credit analysis system.
In the consumer installment market, the main subject is general consumer. We can not apply market –to-market model to value the consumer. The Default model is using to measure the credit risk of individual consumer. The new development approaches, including Generic Algorithms, Artificial Neural Network, Fuzzy theory, and Grey Theory, are seldom used to create personal credit risk model. The special characteristic of consumer installment market—wide product line but shallow product market, limits the applications of above measuring approaches on personal credit scoring system.
We are used to apply subjective method such as 5C or 5P to judge personal credit. In multi-variables analysis the Probit and Logit model are popular in risk measurement model. But their explanation power and estimating coefficient are very similar. We choose only one approach is enough to do empirical research.
The database setup is first priority for credit scoring system creation. We need to include not only the personal factor of consumer but also the basic data of customer, the data setup of credit data, the trading process and the connection with the customer. We need all the database can be worked consistence and cross connection with each other. The database setup, the related operation interface, and the integration of new and old system are the main factors to ensure the fluently usage of credit scoring system. |
| Appears in Collections: | [Graduate Institute & Department of Banking and Finance] Thesis
|
All items in 機構典藏 are protected by copyright, with all rights reserved.
|
