An Efficient SMOTE-Based Deep Learning Model for Heart Attack Prediction

Table 4

Average accuracy, precision, and F1-measure obtained for positive and negative classes after experimentation.

Algorithm	Average precision		Average recall		Average F1-measure
	Positive class	Negative class	Positive class	Negative class	Positive class	Negative class
SMOTE-based AdaBoost	81.6	80.6	83.3	77.1	82	78.1
SMOTE-based bagging	81.4	82.2	84.6	76.2	82.4	78.4
SMOTE-based random forest	83.4	86.5	89.6	78.5	86.3
SMOTE-based KNN	83.9	80.6	82.7	80.6	82.9	80.3
SMOTE-based logistic regression	81.3	82.6	85.3	75.5	82.7	78.1
SMOTE-based naïve Bayes	83.7	79.9	82.9	79.7	83.1	79.6
SMOTE-based support vector machine	82.1	87.3	88.9	76.2	84.7	80.7
SMOTE-based vote	82.1	82.8	85.8	76.3	83.4	78.7
SMOTE-based artificial neural network	97.7	94.1	94.4	97.1	95.9	95.4