Morphometric and radiomics analysis toward the prediction of epilepsy associated with supratentorial low-grade glioma in children

Table 3 Performance of seizure prediction models using radiomics features for pediatric patients with supratentorial glioma

Classifier	Precision	Recall	Specificity	Accuracy	F1-Score	TP	FN	FP	TN	AUC
Coarse Gaussian SVM	0.800	0.870	0.800	0.833	0.833	20	3	5	20	0.859
Medium KNN	0.850	0.739	0.880	0.813	0.791	17	6	3	22	0.851
Linear Discriminant	0.800	0.870	0.800	0.833	0.833	20	3	5	20	0.845
Linear SVM	0.800	0.870	0.800	0.833	0.833	20	3	5	20	0.833
Weighted KNN	0.792	0.826	0.800	0.813	0.809	19	4	5	20	0.833
Cubic KNN	0.762	0.696	0.800	0.750	0.727	16	7	5	20	0.831
Kernel Naive Bayes	0.708	0.739	0.720	0.729	0.723	17	6	7	18	0.828
Cosine KNN	0.800	0.696	0.840	0.771	0.744	16	7	4	21	0.811
Medium Gaussian SVM	0.783	0.783	0.800	0.792	0.783	18	5	5	20	0.809
Fine Gaussian SVM	0.773	0.739	0.800	0.771	0.756	17	6	5	20	0.800
Binary GLM Logistic Regression	0.760	0.826	0.760	0.792	0.792	19	4	6	19	0.793
Gaussian Naive Bayes	0.818	0.391	0.920	0.667	0.529	9	14	2	23	0.772
Quadratic Discriminant	0.688	0.478	0.800	0.646	0.564	11	12	5	20	0.717
Fine Tree	0.682	0.652	0.720	0.688	0.667	15	8	7	18	0.708
Medium Tree	0.682	0.652	0.720	0.688	0.667	15	8	7	18	0.708
Coarse Tree	0.696	0.696	0.720	0.708	0.696	16	7	7	18	0.706
Fine KNN	0.667	0.696	0.680	0.688	0.681	16	7	8	17	0.688
Cubic SVM	0.500	0.044	0.960	0.521	0.080	1	22	1	24	0.115
Quadratic SVM	NaN	0.000	1.000	0.521	NaN	0	23	0	25	0.000
Coarse KNN	NaN	0.000	1.000	0.521	NaN	0	23	0	25	0.000

AUC, area under receiver operating characteristic curve; FN: False Negative; FP: False Positive; GLM: Generalized linear models; KNN: k-nearest neighbors; SVM: Support Vector Machines; TN: True Negative TP: True Positive

ISSN: 1470-7330