Bankalarda Kobi Kredilerini Değerlendirmeye İlişkin Bir Yaklaşım : Yapay Sinir Ağları
Loading...
Date
2007
Authors
Yazici, Mehmet
Journal Title
Journal ISSN
Volume Title
Publisher
Kadir Has Üniversitesi
Open Access Color
OpenAIRE Downloads
OpenAIRE Views
Abstract
Bankaların karsı karsıya oldugu riskler nedeni ile yükümlülüklerini yerine getiremeyecek duruma düsmesini önlemek ve standart uygulamalar gelistirmek amacıyla global düzeyde düzenlemelere gidilmektedir. Bu düzenlemeler ile bankaların ve ekonomik sistemin saglıklı islemesine yönelik muhtemel risklerin izlenmesi ve kontrolünü saglamak üzere olusturacakları iç denetim ve risk yönetim sistemlerine iliskin esas ve usullerin belirlenmesi, tasınan risklere karsılık yeterli sermayenin bulundurulması amaçlanmaktadır. Sorunlu kredilerin olumsuz etkilerinin önüne geçilmesi, kaynakların optimal dagılımı ve verimli kullanılması ancak karsılasılması muhtemel ve öngörülebilir risklerin bugünden tespit edilerek riskin dogru yönetilmesi ve gerekli aksiyonların bugünden alınmasına baglıdır. Bu çalısmanın amacı; risk degerlendirmesinin büyük ölçekli kurumsal firmalara oranla daha zor oldugu KOB'lerde mali basarısızlıgın tahmini ile ilgili olarak alternatif bir yöntem ortaya koymaktır. Yapılmıs olan bu tez çalısmasında, ekonomimizin temelini ve Bankalarımızın son dönemdeki odak noktasını olusturan ve KOB'lerle ilgili basarısızlık tahminlerinde mali verilerin tek basına yeterli olmadıgından hareketle, daha önce basarı ile uygulanmıs olan Diskriminant Analizi, Lojistik Regresyon ve Yapay Sinir Agı yöntemleri ile birer uygulama yapılarak sonuçlar karsılastırılmıs, iyi ve kötü kredi ayrımının Yapay Sinir Agı modeli ile daha basarılı sekilde yapıldıgı sonucuna ulasılmıstır
Global regulations take effect to prevent banks from a position where they are not able to meet their requirements because of risks they encounter and to set standard procedures. With these regulations, in order to achieve sound and reliable banks and economic systems, it is intended to determine internal auditing and risk management principles and procedures that enable banks to monitor and control potential risks and to possess funds that can match risks taken. Preventing micro and macro economic problems, distributing resources optimally, and using resources efficiently may only be possible by determining and managing foreseeable risks and taking necessary actions today. The objective of the current study was to establish a novel method to forecast financial failure in SME?s, where risk evaluation is more complex and difficult compared to large scale organizations. In this dissertation, from a standpoint that financial variables are not adequate to predict financial failure in SME?s, which receive focal attention from banks in recent years, Discriminant Analysis, Logistic Regression, and Artificial Neural Networks methods were utilized separately and compared. Comparing these three techniques, it was found that the discrimination of good and bad credit was best achieved by the Artificial Neural Networks model
Global regulations take effect to prevent banks from a position where they are not able to meet their requirements because of risks they encounter and to set standard procedures. With these regulations, in order to achieve sound and reliable banks and economic systems, it is intended to determine internal auditing and risk management principles and procedures that enable banks to monitor and control potential risks and to possess funds that can match risks taken. Preventing micro and macro economic problems, distributing resources optimally, and using resources efficiently may only be possible by determining and managing foreseeable risks and taking necessary actions today. The objective of the current study was to establish a novel method to forecast financial failure in SME?s, where risk evaluation is more complex and difficult compared to large scale organizations. In this dissertation, from a standpoint that financial variables are not adequate to predict financial failure in SME?s, which receive focal attention from banks in recent years, Discriminant Analysis, Logistic Regression, and Artificial Neural Networks methods were utilized separately and compared. Comparing these three techniques, it was found that the discrimination of good and bad credit was best achieved by the Artificial Neural Networks model
Description
Keywords
Bankacılık, Banking
Turkish CoHE Thesis Center URL
Fields of Science
Citation
WoS Q
Scopus Q
Source
Volume
Issue
Start Page
1
End Page
176