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

Table 10

Results of diagnosing heart disease (Cleveland dataset) by using five machine learning algorithms.

Classifiers	SVM		KNN		Decision Tree		Random Forest		Logistic Regression
Criteria	Training 80%	Testing 20%	Training 80%	Testing 20%	Training 80%	Testing 20%	Training 80%	Testing 20%	Training 80%	Testing 20%
Accuracy (%)	92.56	90.16	87.60	90.16	100	81.97	100	85.25	87.60	88.52
Precision (%)	93.45	90.12	88.82	90.26	100	82.43	100	85.82	88.19	88.56
Recall (%)	92.52	90.45	87.63	90.38	100	82.39	100	85.29	87.44	89.46
F1 score	92.98	90.28	88.23	90.32	100	82.41	100	85.55	87.81	89.00