Table 6 Efficacy and stability of SVM classifier for predicting the differential diagnosis of primary and secondary ICH based on two sketching methods.

From: Machine learning-based CT radiomics model to discriminate the primary and secondary intracranial hemorrhage

Sketch method

Classification

Mean AUC

Precision

Recall

Sensitivity

Specificity

F1-score

VOI sketching images

Primary ICH

0.89 ± 0.05

0.85

0.96

0.96

0.65

0.90

Secondary ICH

0.89 ± 0.05

0.88

0.65

0.65

0.96

0.75

Three-layer sketching images

Primary ICH

0.88 ± 0.05

0.83

0.92

0.92

0.61

0.87

Secondary ICH

0.88 ± 0.05

0.78

0.61

0.61

0.92

0.68

Back to article page