Table 3.

Referenced literature that considered machine-learning-based heart disease diagnosis.

Study	Contributions	Algorithm	Dataset	Data Type	Performance Evaluation
[41]	Predict coronary heart disease	Gaussian NB, Bernoulli NB, and RF	Cleveland dataset	Tabular	Accuracy—85.00%, 85.00% and 75.00%
[42]	Predicting heart diseases	RF, CNN	Cleveland dataset	Tabular	RF (Accuracy—80.327%, Precision—82%, Recall—80%, F1-score—80%), CNN (Accuracy—78.688, Precision—80%, Recall—79%, F1-score—78%)
[43]	Heart disease classification	SVM	Cleveland database	Tabular	Accuracy—73–91%
[44]	Heart disease classification	Back-propagation NN, LR	Cleveland dataset	Tabular	Accuracy (BNN—85.074%, LR—92.58%)
[45]	ECG arrhythmia for heart disease detection	SVM and Cuckoo search optimized NN	Cleveland dataset	Tabular	Accuracy (SVM—94.44%)
[46]	Intelligent scoring system for the prediction of cardiac arrest within 72 h	SVM	Privately ownend	Tabular	Specificity—78.8%, Sensitivity—62.3%, Positive predictive value—10%, Negative predictive value—98.2%
[47]	Automatically identify 5 different categories of heartbeats in ECG signals	CNN	MIT-BIH	Tabular	Accuracy—94% (balance data) Accuracy—89.07% (imbalance data)
[48]	Novel heartbeat recognition method is presented	SVM	MIT-BIH	Tabular	Accuracy—97.77% (imbalance data), Accuracy—97.08% (noise-free ECGs)