Fig 1.
Architecture of our proposed prediction model.
Fig 2.
Ensemble classifier with soft voting.
Fig 3.
Data extraction from KAMIR-NIH dataset.
Table 1.
Hyperparameter tuning for machine learning algorithms.
Table 2.
Variable divisions as categorical and continuous.
Table 3.
Variable divisions as categorical and continuous.
Table 4.
Baseline characteristics of all subjects (N = 11,189).
Fig 4.
Top 10 primary prognostic factors and their feature importance for. (a) RF; (b) ET; (c) GBM.
Table 5.
Performance measures for machine learning models on complete dataset (%).
Table 6.
Performance measures for machine learning models on STEMI dataset (%).
Table 7.
Performance measures for machine learning models on NSTEMI dataset (%).
Table 8.
Overall evaluation results for prediction of MACE on complete dataset (%).
Table 9.
Evaluation results for prediction of MACE on STEMI dataset (%).
Table 10.
Evaluation results for prediction of MACE on NSTEMI dataset (%).
Fig 5.
Normalized confusion matrices for proposed soft voting ensemble classifier on. (a) complete dataset; (b) STEMI dataset; (c) NSTEMI dataset.
Table 11.
Unpaired t-test results for accuracy between STEMI and NSTEMI groups.