http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Application research of consumer credit Score model based on SVM and DEA
Wensheng Dai,Chi-Jie Lu 인하대학교 정석물류통상연구원 2009 인하대학교 정석물류통상연구원 학술대회 Vol.2009 No.10
Clustering and Classification are the most popular techniques in credit scoring. Most of the hybrid models are lack of revision abilities. To overcome this limitation, a two stages credit scoring model using SVM and DEA is proposed in this paper. Firstly constructing a SVM model and classifying all customs to two groups which are efficient and inefficient group, the performing DEA for those lower efficiency customers and proposing the way to improve their efficiency. This study also performs an empirical research based on the credit card database of a bank. The results show that the SVM has great ability to predict the efficiency and combined model can provide an indeed improvement for bank to improve the efficiency of non-efficient customer.
Combining ICA and SVR in Times Series Predication
Wensheng Dai,Jui-Yu Wu,Chi-Jie Lu 인하대학교 정석물류통상연구원 2009 인하대학교 정석물류통상연구원 학술대회 Vol.2009 No.10
In this paper, a time series prediction approach by combing independent component analysis (ICA) and support vector regression (SVR)is proposed ICA is a novel statistical signal processing technique that was originally proposed to find the latent source signals from observed mixture signal without knowing any prior knowledge of the mixing mechanism. SVR is and artificial intelligence forecasting technique and has been widely applied in time series prediction problems. The proposed approach First uers ICA to the forecasting variables for generating the independent components(ICs). After identifying and removing the ICs containing the noise, the rest of the ICs are then used to reconstruct the forecasting variables which contain less noise. The SVR then uses the denoised forecasting variables to build the forecausting model. in order to evaluaate the performance of the proposed approach the TAIEX(Taiwan Stock Exchange Capitalization weighted Steock index) closing cash index is usde as the illusrtative example. Experimental results show that the proposed model outperforms the SVR model with mon ?filtered forecasting variables and random walk model