Table 4. Variables and metrics used in evaluation of model performance.

Metrics/Variable	Definition
n	Total number of data sets
c1true	Number of data sets for infants with IVH
c2true	Number of data sets for infants without IVH
c1method	Number of data sets classified by the ML-method as of infants with IVH
c2method	Number of data sets classified by the ML-method as of infants without IVH
n1	c1true × c1method/n
n2	c2true × c2method/n
Expected accuracy (EA)	(n1 + n2)/n
Observed accuracy (OA)	Number of correctly identified data sets/n
Cohen’s kappa	(OA-EA)/(1-EA)
Mean accuracy	Mean value of OA
n1true	Number of correctly identified data sets for infants with IVH
n2true	Number of correctly identified data sets for infants without IVH
Sensitivity	n1true/c1true
Specificity	n2true/c2true