DocumentCode :
3282513
Title :
Applying Feature Selection Combination-Based Rough Set Classifiers to Forecast Credit Rating Status
Author :
You-Shyang Chen ; Ching-Hsue Cheng ; Da-Ren Chen ; Wei-Yu Chen
Author_Institution :
Dept. of Inf. Manage., Hwa Hsia Inst. of Technol., New Taipei, Taiwan
fYear :
2012
fDate :
25-28 Aug. 2012
Firstpage :
425
Lastpage :
428
Abstract :
When banks experience financial scandals or insolvency, panic typically ensues and, in the worst case, leads to systemic banking crises, they are clearly vital to financial market stability, particularly large bank. Therefore, developing an indicator that represents the financial status and operational competence of Asian banks is urgently needed for parties interested in investing in Asia. This study proposes a stepped model that first organizes random forest (RF) and reducts and core of rough set exploration system (RSES) to construct various combinations of extracted key attributes for reducing data dimensions. Accordingly, the rough set LEM2 algorithm is employed as evaluation method to test the various combinations. for verification, a practical dataset comprising 1,327 samples is collected from the BANKSCOPE database, comprising Asian banks covered the period 1993¡V2007. the experimental results indicate that the proposed model outperforms the listing models in terms of accuracy and its standard deviation.
Keywords :
financial management; rough set theory; Asian banks; BANKSCOPE database; LEM2 algorithm; RF; RSES; data dimensions; feature selection combination based rough set classifier application; financial market stability; financial scandals; financial status; forecast credit rating status; key attributes extraction; operational competence; random forest; rough set exploration system; Accuracy; Banking; Data mining; Data models; Investments; Predictive models; Support vector machines; credit ratings; feature selection; random forest; rough set theory;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Genetic and Evolutionary Computing (ICGEC), 2012 Sixth International Conference on
Conference_Location :
Kitakushu
Print_ISBN :
978-1-4673-2138-9
Type :
conf
DOI :
10.1109/ICGEC.2012.67
Filename :
6457129
Link To Document :
بازگشت