TY - GEN
T1 - Entropy multi-hyperplane credit scoring model
AU - Laesanklang, Wasakorn
AU - Sinapiromsaran, Krung
AU - Intiyot, Boonyarit
PY - 2010
Y1 - 2010
N2 - Entropy multi-hyperplane credit scoring model is a decision model that classifies applicants into payers or defaulters by optimizing the classification cost using multiple hyperplanes based on entropy order. In the first stage, the model uses a pair of hyperplanes composed of half of the attributes which are ordered increasingly by the entropy. The hyperplanes divide the applicants into 3 groups, namely payers, defaulters and the unclassified. From the unclassified group, the model uses another pair of hyperplanes which are composed of additional half of the rest of attributes based on the previous entropy order. The additional hyperplanes divide the unclassified group in the first stage into another 3 group namely payers in 2nd stage, defaulters in 2 nd stage and the unclassified in 2nd stage. In the final stage, the multi-dimensional hyperplanes created from all attributes are used to divide the loaners into two groups : payers and defaulters. In this paper, a mixed-integer programming model for entropy multi-hyperplane credit scoring model is developed to minimize the cost of misclassification errors. The experiment shows that our model has more accuracy than a two-stage least cost credit scoring model and uses less computational iterations than a multi-hyperplane credit scoring model. Moreover, the new model exhibits comparable result with classification tree, neural network, support vector machine, linear discriminant analysis and CART.
AB - Entropy multi-hyperplane credit scoring model is a decision model that classifies applicants into payers or defaulters by optimizing the classification cost using multiple hyperplanes based on entropy order. In the first stage, the model uses a pair of hyperplanes composed of half of the attributes which are ordered increasingly by the entropy. The hyperplanes divide the applicants into 3 groups, namely payers, defaulters and the unclassified. From the unclassified group, the model uses another pair of hyperplanes which are composed of additional half of the rest of attributes based on the previous entropy order. The additional hyperplanes divide the unclassified group in the first stage into another 3 group namely payers in 2nd stage, defaulters in 2 nd stage and the unclassified in 2nd stage. In the final stage, the multi-dimensional hyperplanes created from all attributes are used to divide the loaners into two groups : payers and defaulters. In this paper, a mixed-integer programming model for entropy multi-hyperplane credit scoring model is developed to minimize the cost of misclassification errors. The experiment shows that our model has more accuracy than a two-stage least cost credit scoring model and uses less computational iterations than a multi-hyperplane credit scoring model. Moreover, the new model exhibits comparable result with classification tree, neural network, support vector machine, linear discriminant analysis and CART.
KW - A two-stage least cost credit scoring model
KW - Credit scoring
KW - Decision model
KW - Entropy
KW - Hyperplane
KW - Mixed-integer programming
UR - http://www.scopus.com/inward/record.url?scp=77955287849&partnerID=8YFLogxK
U2 - 10.1109/ICFTE.2010.5499418
DO - 10.1109/ICFTE.2010.5499418
M3 - Conference contribution
AN - SCOPUS:77955287849
SN - 9781424477586
T3 - 2010 International Conference on Financial Theory and Engineering, ICFTE 2010
SP - 91
EP - 94
BT - 2010 International Conference on Financial Theory and Engineering, ICFTE 2010
T2 - 2010 International Conference on Financial Theory and Engineering, ICFTE 2010
Y2 - 18 June 2010 through 20 June 2010
ER -