SpringerPlus

Table 7 LASSO–CART model—the fivefold cross validation results

From: Improving the prediction of going concern of Taiwanese listed companies using a hybrid of LASSO with data mining techniques

Subset	Training set					Testing set
Subset	Predicted group		Hit ratio (%)	Type I error (%)	Type II error (%)	Predicted group		Hit ratio (%)	Type I error (%)	Type II error (%)
1	66	6	93.75	8.33	0.00	68	4	93.75	5.56	8.33
	0	24				2	22
2	70	2	93.75	2.78	16.67	57	15	86.46	20.83	16.67
	4	20				4	20
3	67	5	92.71	6.94	8.33	65	7	90.62	9.72	20.83
	2	22				5	19
4	69	3	92.71	4.17	16.67	60	12	83.33	16.67	12.50
	4	20				3	21
5	72	0	94.79	0.00	20.83	61	11	89.58	15.28	12.50
	5	19				3	21
Avg.			93.54	4.44	12.50			88.75	13.61	14.17

Back to article page