Improving an Intelligent Detection System for Coronary Heart Disease Using a Two-Tier Classifier Ensemble

Table 5

The selected features obtained from PSO based feature selection for each dataset.

Dataset	# selected features	Feature name
Z-Alizadeh Sani	27	Age, hypertension, airway disease, thyroid disease, congestive heart failure, dyslipidemia, blood pressure, systolic murmur, diastolic murmur, typical chest pain, dyspnea, atypical, nonanginal, low threshold angina, ST elevation, T inversion, poor R progression, fasting blood sugar, LDL, HDL, blood urea nitrogen, erythrocyte sedimentation rate, white blood cell, neutrophil, ejection fraction, region with regional wall motion abnormality, and valvular heart disease.
Statlog	8	Gender/sex, chest pain type, resting electrocardiographic results, maximum heart rate achieved, exercise induced angina, ST depression, number of major vessels, and thallium stress test result.
Cleveland	7	Chest pain type, resting electrocardiographic results, maximum heart rate achieved, exercise induced angina, oldpeak, number of major vessels, and thallium stress test result.
Hungarian	6	Gender/sex, chest pain type, heart rate, old peak, slope, and number of major vessels.