TY - GEN
T1 - Credit scoring based on eigencredits and SVDD
AU - Wei, Haizhou
AU - Li, Jianwu
PY - 2011
Y1 - 2011
N2 - Credit risk evaluation is an active research topic in financial risk management, and credit scoring is an important analytical technique in credit risk evaluation. In this paper, a new two-stage method is introduced to perform credit scoring. Eigencredits are firstly constructed based on creditworthy examples through principal component analysis to extract the principal features of creditworthy data. Then, support vector domain description (SVDD) is further used to describe creditworthy examples. Preliminary experiments based on two real data sets from UCI repository show the effectiveness of the proposed method.
AB - Credit risk evaluation is an active research topic in financial risk management, and credit scoring is an important analytical technique in credit risk evaluation. In this paper, a new two-stage method is introduced to perform credit scoring. Eigencredits are firstly constructed based on creditworthy examples through principal component analysis to extract the principal features of creditworthy data. Then, support vector domain description (SVDD) is further used to describe creditworthy examples. Preliminary experiments based on two real data sets from UCI repository show the effectiveness of the proposed method.
KW - Credit scoring
KW - Eigencredits
KW - Support vector domain description (SVDD)
UR - http://www.scopus.com/inward/record.url?scp=80052351017&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-23220-6_5
DO - 10.1007/978-3-642-23220-6_5
M3 - Conference contribution
AN - SCOPUS:80052351017
SN - 9783642232190
T3 - Communications in Computer and Information Science
SP - 32
EP - 40
BT - Applied Informatics and Communication - International Conference, ICAIC 2011, Proceedings
T2 - 2011 International Conference on Applied Informatics and Communication, ICAIC 2011
Y2 - 20 August 2011 through 21 August 2011
ER -