Loading...
|
Please use this identifier to cite or link to this item:
https://nccur.lib.nccu.edu.tw/handle/140.119/78861
|
| Title: | 應用資料採礦技術建置台灣中小企業電子業信用評等模型 |
| Other Titles: | Application Data Mining Technology and the Establishmentof Taiwan`s Electronics Industry Credit Rating-Sized Enterprises, etc Models |
| Authors: | Chen, Guan-Yu;Su, Chih-Hsiung
陳冠宇 |
| Contributors: | 統計系 |
| Keywords: | 新巴塞爾協定;資料採礦;羅吉斯迴歸;信用評等;信用轉移矩陣
The New Basel Accord;Data Mining;Luo JI Regression;Credit Rating;Credit Transfer Matrices |
| Date: | 2009-10 |
| Issue Date: | 2015-10-06 14:51:37 (UTC+8) |
| Abstract: | 全球化潮流方興未艾，基於與國際接軌目標，我國金融業自2006年起實施新巴塞爾資本協定，期於現今日新月異金融環境中以全球一致性的銀行管理方法及制度落實其精神。實施新巴塞爾協定後，首當其衝者便是台灣產業發展主體－中小企業。以信用風險中資本計提為例，中小企業不若大型企業體質健全，且財務透明度亦為人詬病，相對提升金融機構授信風險，進而導致中小企業融資授信審查趨於嚴格與保守，中小企業融資難度與成本皆大幅增加。有鑑於此，本研究以中小企業中電子業為主要研究對象，採資料採礦流程進行信用評等模型建置。為求配適最佳違約機率模型，分別以不同精細抽樣比例逐一配適羅吉斯迴歸、類神經網路及分類迴歸樹等統計模型，經評估後篩選出羅吉斯迴歸模型建置信用評等系統。再者，為確認模型與信用評等系統建置適當，係遵循新巴塞爾協定相關規範進行各項測試及驗證，結果顯示模型於樣本外資料測試表現良好，信用評等系統亦通過正確性分析、等級區隔同質性檢定及穩定度分析等驗證準則，冀能提供金融機構一套有效且精簡的信用管理機制，建立與中小企業間資訊對稱管道，於兩造雙方取得互利平衡，防範危機於未然。
Tide of globalization in the ascendant, target based on the international practice, our finance implemented since 2006 the new Basel Capital Accord, period on technology by a global financial environment of consistency in Bank management methods and systems to implement its spirit. After implementing the new Basel Capital Accord, the brunt will be the subject of Taiwan`s industrial development-SMEs. To credit risk capital recognized as an example, if a large enterprise-sized enterprises do not sound, and physical fitness financial transparency also criticized, relative to enhance financial institutions credit risk, resulting in the financing of the credit review tends to be more strictly and conservative, difficulty and cost of financing of both the young has increased substantially. In view, this study to the electronics industry in the medium-sized enterprises as a key study object, mining data mining process carried out on the establishment of credit rating model. For the sake of matching the best odds model, respectively, in breach of contract to different fine sampling ratio of one by one matching Luo JI regression, class of neural networks and classification and regression trees etc statistical model, after assessment to filter for the establishment of Luo JI regression model credit rating systems. Furthermore, to confirm that the model and credit rating system and the establishment of appropriate, lines follow the new Basel Capital Accord-related norms for testing and validation, the results show model to the sample data test performed well, credit rating system also endorsed the correctness of analysis, grade tile homogeneous verification and validation criteria such as stability analysis, financial institutions can provide an effective and streamline the credit management mechanism, establishment and SMEs information symmetry pipeline, on both sides acquire beneficial and balanced, preventive crisis in the first place. |
| Relation: | Journal of Data Analysis, 4(5), 23-53 |
| Data Type: | article |
| Appears in Collections: | [經濟學系] 期刊論文
|
Files in This Item:
| File |
Description |
Size | Format | |
|---|
| 23-53.pdf | | 2032Kb | Adobe PDF2 | 391 | View/Open |
|
All items in 政大典藏 are protected by copyright, with all rights reserved.
|
