Texture analysis of pulmonary parenchymateous changes related to pulmonary thromboembolism in dogs – a novel approach using quantitative methods

Table 3 Sensitivity, specificity, confidence intervals and classification error for the performance of PLS-DA and SVM

	Sensitivity	CI lower bound	CI upper bound	Specificity	CI lower bound	CI upper bound	Classification error
PLSDA
Healthy (CA)	0,9734	0,9597	0,9825	0,979	0,9687	0,986	0,02335
Healthy (CV)	0,9443	0,9261	0,9583	0,9616	0,9485	0,9715	0,04565
Diseased with PTE (CA)	0,5912	0,5344	0,6457	0,8558	0,8377	0,8722	0,18,577
Diseased with PTE (CV)	0,3041	0,2544	0,3587	0,8029	0,7826	0,8217	0,27,548
Diseased without PTE (CA)	0,7043	0,6717	0,7349	0,884	0,83,636	0,9017	0,19,214
Diseased without PTE (CV)	0,589	0,5545	0,6226	0,7947	0,7696	0,8176	0,29,246
SVM
Healthy (CA)	1	0,9925	1	1	0,9925	1	0
Healthy (CV)	0,9835	0,9721	0,09904	0,9963	0,9906	0,9986	0,00902
Diseased with PTE (CA)	0,9189	0,8822	0,9449	0,9836	0,9761	0,9888	0,02654
Diseased with PTE (CV)	0,3885	0,3348	0,4451	0,8967	0,8808	0,9107	0,18,312
Diseased without PTE (CA)	0,9674	0,9527	0,9777	0,9779	0,9673	0,9851	0,02654
Diseased without PTE (CV)	0,7895	0,7598	0,8163	0,8214	0,7975	0,843	0,19,214

ISSN: 1746-6148