Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/27151
標題: 農會信用部財務危機預測模型之研究-模糊類神經網路系統之應用
The Forecasting Model of Financial Crisis for the Credit Departments of Farmers' Associations in Taiwan - the Application of Neural Fuzzy Network System
作者: 施麗玉
Shih, Lee-Yu
關鍵字: the Credit Department of Farmers' Association;農會信用部;financial crisis;neural fuzzy network;財務危機;模糊類神經網路
出版社: 農業經濟學系
摘要: 
農會信用部在農業金融體系中扮演舉足輕重的重要角色,其經營管理良窳對整體金融及經濟體系之安定與繁榮具有關鍵性的影響。當前農會信用部財務狀況惡化,而金融監理制度未能對財務危機農會信用部妥善預警與輔導,不但影響農會持續經營與農業金融體系功能完整性且將危及整體金融紀律,有建立財務危機預測之必要。農會信用部之經營績效因受諸多內外部因素影響而呈現評估不易之模糊現象,如經營風險大小、資本適足程度、財務危機程度等,均無法以精確之數學參數加以模式化。以往之研究方法以類神經網路模型較佳,但仍未脫離傳統嚴格二分法之評判誤差。本文採用自我建構模糊類神經推論網路系統預測農會信用部財務危機等級,輸入變數乃參考CAMELS評估指標,結合金融機構財務危機相關理論與農會信用部經營特性選取24個指標,輸出變數以因素分析方法劃分268個農會信用部民國87至89年度評等等級,實證結果得到相當不錯之準確率,應可供政府建立農會信用部財務危機預警模型之參考。由評等結果發現,民國90年9月被政府強制合併的22家樣本農會信用部,在87年已落入評等第D、E級者有18家,88年有21家,可見假設農會信用部之財務資料揭露無誤的話,22家中有18家早於被整頓之前3年已有財務危機,餘4家則亦在前2年現出危機跡象,應可提早採取適當監理機制,避免財務危機的發生及被強制合併之後果。

The credit departments of Framers' Associations have been playing a very important role in the agricultural finance system in Taiwan. Currently, some credit departments are in serious financial problems, meanwhile, without efficient and effective financial inspections and operational assistance under financial investigation mechanism. As a result, a more effective financial forecasting model should be built for the reduction of operational crisis of credit departments. The above unsound situation was resulted from the imprecise fuzzy phenomenon on evaluating the operational performance of credit departments such as the scale of operational risks, the level of capital adequacy, the degree of financial crisis, etc. Mathematical parameters are not able to accurately explore these fuzzy phenomenon. At the same time, past researchers used Neural Network System which adopted only hard dichotomous classification method to create certain degree of forecast errors. This research adopted Self-Constructing Neural Fuzzy Inference Network System to forecast different A to E downward ranks of financial crisis for credit departments. The parameters of CAMELS are used as inputs and rank variables calculated from Factor Analysis are as outputs. Totally, 268 credit department samples during 1998 and 2000 are analyzed with excellent forecasted output results. The results showed that the liquidated 22 credit departments in the September of 2001 under government enforcement had 18 and 21 credit departments as D and E ranks in 1998 and 1999, respectively. If this research forecasted of financial crisis performed in 1998, then 18 and 4 out of the 22 liquidated credit departments would have been dealt with proper supervision mechanism three and two years ago, respectively. The solvency crisis and enforced consolidation would not happen.
URI: http://hdl.handle.net/11455/27151
Appears in Collections:應用經濟學系

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.