Score and Correlation Coefficient-Based Feature Selection for Predicting Heart Failure Diagnosis by Using Machine Learning Algorithms

Table 2

Heart failure features with metrics from the Allied Hospital dataset.

Features	Explanation	Range	Measurement
Age	Age of patient in year	[40, ..., 95]	Year
Anaemia	1 = haematocrit levels lower than 36%	0, 1	Boolean
	0 = haematocrit levels higher than 36%
High blood pressure	1 = patient has hypertension	0, 1	Boolean
	0 = patient has no hypertension
Creatinine phosphokinase	Level of CPK in blood	[23, ..., 7861]	mcg/L
Diabetes	1 = patient has diabetes	0, 1	Boolean
	0 = patient has no diabetes
Sex	1 = male	0, 1	Boolean
	0 = female
Platelets	Blood platelets	[25.01, ..., 850.00]	Kiloplatelets/mL
Serum creatinine	Level of creatinine in blood	mg/dL	[0.50, ..., 9.40]
Serum sodium	Level of sodium in blood	mEq/L	[114, ..., 148]
Smoking	1 = patient smokes	0, 1	Boolean
	0 = patient does not smoke
Time	Periodic follow-up of patient	Days	[4,..., 285]
Death event (target)	1 = patient died during follow-up	0, 1	Boolean
	0 = patient did not die during follow-up

mcg/L refers to micrograms per litre. mL refers to microlitre. mEq/L refers to milliequivalents per litre.