Table 5. Predicted risks and reclassification of STEMI patient’s mortality between ML best model (SVMvarImp-SBE-SVM) and TIMI risk score for in-hospital, 30-days, and 1-year on 30% validation dataset.

In-hospital
Individuals with events (n = 101)
		Number of individuals		Reclassified		Net correctly reclassified (%)
		Low risk	High risk	Increased risk	Decreased risk	18.81
		Machine Learning		23	4
	TIMI Score
	Low risk	13	23
	High risk	4	61
Individuals without events (n = 1788)
		Machine Learning		116	144	1.57
	TIMI Score
	Low risk	1375	116
	High risk	144	153
Net reclassification index (NRI)						20.38
p-value						<0.0001
30 days
Individuals with events (n = 75)
		Number of individuals		Reclassified		Net correctly reclassified (%)
		Low risk	High risk	Increased risk	Decreased risk	20.00
		Machine Learning		17	2
	TIMI Score
	Low risk	11	17
	High risk	2	45
Individuals without events (n = 863)
		Machine Learning		66	53	-1.51
	TIMI Score
	Low risk	652	66
	High risk	53	92
Net reclassification index (NRI)						18.49
p-value						<0.0001
1-year
Individuals with events (n = 126)
		Number of individuals		Reclassified		Net correctly reclassified (%)
		Low risk	High risk	Increased risk	Decreased risk	23.81
		Machine Learning		44	14
	TIMI Score
	Low risk	21	44
	High risk	14	47
Individuals without events (n = 754)
		Machine Learning		108	36	-9.55
	TIMI Score
	Low risk	523	108
	High risk	36	87
Net reclassification index (NRI)						14.26
p-value						<0.0001