Evaluating the Risk of Metabolic Syndrome Based on an Artificial Intelligence Model

Table 1

Calculation of the sensitivity and specificity for a specific cut-off point of the predicted probability 𝑃.

Model result	Reality
	Metabolic syndrome	No MetS	Total cases
Metabolic syndrome	𝑎 (true positive)	𝑏 (false positive)	𝑎+𝑏
No MetS	𝑐 (false negative)	𝑑 (true negative)	𝑐+𝑑
Total cases	𝑎+𝑐	𝑏+𝑑

Sensitivity =𝑎/(𝑎+𝑐)= true positive rate.
Specificity =𝑑/(𝑏+𝑑)= true negative rate.
l − specificity =1−[𝑑/(𝑏+𝑑)]=𝑏/(𝑏+𝑑)= false positive rate.
“𝑎” cases have a 𝑃 > cut-off point, which have an observed feather as well.
“𝑏” cases have a 𝑃 > cut-off point, although no feather was observed in reality.
“𝑐” cases have a 𝑃 < cut-off point, although feather was observed in reality.
“𝑑” cases have a 𝑃 < cut-off point, and for these cases no feather was observed.