|
| Authors, year, country | Aim of study | Data and features | Sample size | ML method and algorithms | Performance | Validation technique | Detail |
|
| Abdar et al. 2019, Poland [16] | Accurate diagnosis | Alizadeh dataset: demographic; symptom, examination, ECG, laboratory, echo | 303 | C-SVC, NU SVC, linear SVM | F1score = 91.51 Acc = 93.08 | 10-fold | One hot encoding, genetic algorithm, genetic optimizer |
|
| Gupta et al. 2019, Canada [17] | Estimating the risk of CAD | Z-Alizadeh Sani (demographic, health history, medical procedure features) | 303 | BN (Bayesian network) | AUC = (0.93 + 0.04) | 10-fold | LR, SVM, ANN graphical reasoning introduces |
|
| Joloudari et al. 2020, Iran [18] | CAD diagnosis | Z-Alizadeh Sani dataset | 303 | DT (Decision tree) | AUC = 91.47 | 10-fold | RTS (Random tree), SVM, DT |
|
| Tama et al. 2020, S. Korea [19] | Detection CHD | 5 dataset (Z-Alizadeh Sani, statlog, cleveland, Hungarian) | 303 | Two-tier ensemble (GBM, GXboost, RF) | Proposed AUC > other ensemble and individual models | 10-fold | Random forest (RF), gradient boosting. Correlation-based feature |
|
| Iong et al. 2021, Taiwan. [20] | Early prediction of CAD | 7 feature (demographic and medical history) | NM | SVM with pooling layer | SVM | NM | SVN NB |
|
| Chen et al. 2020, China [21] | Detection of CAD | 1163 variables (morphological) | | Polynomial SVM with grid search optimization | Acc = 100% | 10-fold | LR, DT, LDA, KNN, ANN.SVM |
|
| Zhang et al. 2020, China [22] | Detection of CAD | Holter monitoring, echocardiography (ECHO), and biomarker levels (BIO) | 62 | Holter model | Sen = 96.67% Spe = 96.67% Acc = 96.64% | 5-fold | Random forest, and SVM. Bioexamination reach the best result. |
|
| Ricciardi et al. 2020, Italy [23] | Prediction of CAD | 22 features (laboratory and medical history) | 10,265 | LDA and PCA | Acc = 84.5 and 86.0 Spe > 97% Sen > 66% | 10-fold | PCA and LDA for feature extraction PostgreSQL, a DBMS |
|
| Pattarabanjird et al., 2020, USA [24] | Prediction of CAD severity | Demographics and laboratory | 481 | NN (ID3 rs11574) | AUC = 72% to 84% | NM | Crf, ID3 |
|
