Financial Distress Prediction of Chinese Listed Companies Using the Combination of Optimization Model and Convolutional Neural Network

Table 9

Prediction results of different models.

Years	Criterion	NN	CART	SVM	LR	Ours
2012–2016	Accuracy	92.66%	80.39%	92.37%	88.59%	94.5%
	KS	0.8164	0.5987	0.8087	0.738	0.9857
	AUC	0.8577	0.6126	0.8517	0.7786	0.9974
2012–2015	Accuracy	72.98%	54.38%	86.90%	85.10%	88.52%
	KS	0.3771	0.5084	0.6415	0.6025	0.7774
	AUC	0.513	0.2908	0.7539	0.7234	0.9548
2016	Accuracy	83.05%	82.15%	91.09%	88.92%	N/A
	KS	0.5769	0.6266	0.7689	0.7224	N/A
	AUC	0.6856	0.6521	0.8286	0.7881	N/A
2015	Accuracy	85.52%	77.92%	84.09%	86.38%	N/A
	KS	0.5906	0.5401	0.6016	0.6526	N/A
	AUC	0.731	0.569	0.7025	0.743	N/A
2014	Accuracy	68.82%	70.17%	76.12%	73.54%	N/A
	KS	0.3451	0.387	0.2983	0.2884	N/A
	AUC	0.4658	0.4352	0.5793	0.5383	N/A
2013	Accuracy	76.38%	76.12%	79.42%	71.48%	N/A
	KS	0.3827	0.4603	0.3898	0.2662	N/A
	AUC	0.5789	0.5605	0.6306	0.508	N/A
2012	Accuracy	72.46%	71.67%	72.19%	64.71%	N/A
	KS	0.3346	0.3881	0.3025	0.3435	N/A
	AUC	0.5196	0.4799	0.5201	0.4175	N/A