Lipids in Health and Disease

Table 4 Performance indices of the DT model for Dyslipidemia

From: Predicting the 10-year incidence of dyslipidemia based on novel anthropometric indices, using data mining

A: Performance indices of the DT adjusted model for variables without confounding factors
(a) Training (n = 1379)			(b) Testing (n = 345)
Actual	Predicted Count		Actual	Predicted Count
	Non- Dyslipidemia	Dyslipidemia		Non-Dyslipidemia	Dyslipidemia
Non-Dyslipidemia	171	367	Non-Dyslipidemia	32	107
Dyslipidemia	106	767	Dyslipidemia	31	192
Specificity = 87%	AUC = 65%		Specificity = 59%		AUC = 67%
Accuracy = 66%	Precision = 67%		Accuracy = 48%		Precision = 64%
B: Performance indices of the DT adjusted model for variables with confounding factors
(a) Training (n = 1379)			(b) Testing (n = 345)
Actual	Predicted Count		Actual	Predicted Count
	Non-Dyslipidemia	Dyslipidemia		Non-Dyslipidemia	Dyslipidemia
Non-Dyslipidemia	231	322	Non-Dyslipidemia	48	78
Dyslipidemia	151	712	Dyslipidemia	49	185
Specificity = 82%	AUC = 65%		Specificity = 70%		AUC = 67%
Accuracy = 66%	Precision = 68%		Accuracy = 64%		Precision = 79%

Back to article page

ISSN: 1476-511X

Contact us

General enquiries: journalsubmissions@springernature.com