A high-dimensionality-trait-driven learning paradigm for high dimensional credit classification

Financial Innovation

Table 7 Robustness analysis of classification without feature extraction in Kaggle dataset when feature dimensions ≥ sample size

	Single linear classifier					Single nonlinear classifier
	Total	TP	TN	AUC		Total	TP	TN	AUC
Results of 10 experiments	0.6190	0.6538	0.5849	0.6636	Results of 10 experiments	0.6000	0.6827	0.5189	0.6463
	0.6381	0.6887	0.5865	0.6910		0.6488	0.6533	0.6442	0.6919
	0.5667	0.5802	0.5529	0.6020		0.5750	0.4976	0.6538	0.6044
	0.6143	0.5872	0.6436	0.6590		0.5940	0.5734	0.6163	0.6268
	0.6429	0.6228	0.6667	0.6894		0.6119	0.6053	0.6198	0.6483
	0.6357	0.6250	0.6462	0.6828		0.6107	0.6250	0.5967	0.6454
	0.6548	0.6905	0.6190	0.6806		0.5917	0.5738	0.6095	0.6378
	0.5762	0.6237	0.5354	0.6201		0.5845	0.6134	0.5597	0.6320
	0.6381	0.7050	0.5773	0.6842		0.6179	0.7600	0.4886	0.6593
	0.6190	0.6582	0.5848	0.6562		0.5560	0.6224	0.4978	0.5887
Average	0.6205	0.6435	0.5997	0.6629	Average	0.5990	0.6207	0.5805	0.6381
Standard deviation	0.0287	0.0431	0.0427	0.0303	Standard deviation	0.0255	0.0701	0.0604	0.0285

	Linear ensemble classifier					Nonlinear ensemble classifier
	Total	TP	TN	AUC		Total	TP	TN	AUC
Results of 10 experiments	0.6357	0.6394	0.6321	0.6644	Results of 10 experiments	0.6190	0.6563	0.5825	0.6595
	0.6452	0.6887	0.6010	0.6830		0.6417	0.6132	0.6707	0.6748
	0.5952	0.5849	0.6058	0.6142		0.5690	0.4906	0.6490	0.5846
	0.6095	0.6101	0.6089	0.6497		0.5881	0.5321	0.6485	0.6297
	0.6333	0.6360	0.6302	0.6810		0.5702	0.5044	0.6484	0.6146
	0.6429	0.6538	0.6321	0.6887		0.6167	0.6995	0.5354	0.6379
	0.6214	0.6333	0.6095	0.6593		0.6143	0.5786	0.6500	0.6590
	0.5952	0.6237	0.5708	0.6203		0.5750	0.6289	0.5288	0.6133
	0.6357	0.6900	0.5864	0.6959		0.5595	0.7500	0.3864	0.6296
	0.6000	0.6020	0.5982	0.6355		0.5845	0.6327	0.5424	0.6150
Average	0.6214	0.6362	0.6075	0.6592	Average	0.5938	0.6086	0.5842	0.6318
Standard deviation	0.0199	0.0343	0.0202	0.0288	Standard deviation	0.0273	0.0837	0.0886	0.0270