From: A high-dimensionality-trait-driven learning paradigm for high dimensional credit classification
Feature extraction | Classifer | Total | TP | TN | AUC | |
---|---|---|---|---|---|---|
PCA-based feature extraction (Non-feature extraction) | Single linear classifier | LogR | 0.6884 (0.6803) | 0.7030 (0.6865) | 0.6737 (0.6743) | 0.7574 (0.7528) |
LDA | 0.6897 (0.6011) | 0.7009 (0.6360) | 0.6785 (0.5658) | 0.7592 (0.6351) | ||
Single nonlinear classifier | KNN | 0.6360 (0.6144) | 0.8205 (0.7434) | 0.4485 (0.4835) | 0.7015 (0.6653) | |
SVM | 0.7112 (0.7012) | 0.7411 (0.7048) | 0.6809 (0.6976) | 0.7822 (0.7707) | ||
BPNN | 0.6771 (0.6537) | 0.6731 (0.6515) | 0.6812 (0.6568) | 0.7468 (0.7131) | ||
CART | 0.5676 (0.6382) | 0.5686 (0.6402) | 0.5667 (0.6364) | 0.5636 (0.6359) | ||
Linear ensemble classifier | LogR Bagging | 0.6791 (0.6861) | 0.6948 (0.7079) | 0.6631 (0.6639) | 0.7447 (0.7579) | |
LDA Bagging | 0.6793 (0.6003) | 0.6917 (0.6387) | 0.6670 (0.5614) | 0.7441 (0.6363) | ||
Nonlinear ensemble classifier | KNN Bagging | 0.6120 (0.6331) | 0.8307 (0.7305) | 0.3892 (0.5345) | 0.6702 (0.6876) | |
SVM Bagging | 0.6972 (0.6813) | 0.7290 (0.6243) | 0.6650 (0.7396) | 0.7652 (0.7459) | ||
BPNN Bagging | 0.6747 (0.6686) | 0.6825 (0.6618) | 0.6668 (0.6752) | 0.7340 (0.7327) | ||
CART Bagging | 0.5991 (0.6372) | 0.6313 (0.6138) | 0.5662 (0.6610) | 0.5908 (0.6330) |