| Pei et al. (2019) |
China |
4,205 |
Binary |
J48 has the best performance (accuracy = 0.9503, precision = 0.950, recall = 0.950, F-measure = 0.948, and AUC = 0.964) |
| Wu et al. (2018) |
USA |
768 |
Binary |
The proposed model attained a 3.04% higher prediction accuracy than those of other studies |
| Talaei-Khoei & Wilson (2018) |
Australia |
10,911 |
Binary |
The performance of different learners depends on both period and purpose of prediction |
| Upadhyaya et al. (2017) |
USA |
4,208 |
Binary |
The proposed algorithm performed well with a 99.70% sensitivity and a 99.97% specificity |
| Nilashi et al. (2017) |
USA |
768 |
Binary |
The proposed method remarkably improves the accuracy of prediction in relation to prior methods |
| Maniruzzaman et al. (2017) |
USA |
768 |
Binary |
The performance of Gaussian process classification are better than other methods with accuracy = 81.97%, sensitivity = 91.79%, positive predictive value = 84.91%, and negative predictive value = 62.50% |
| Kagawa et al. (2017) |
Japan |
104,522 |
Binary |
The proposed phenotyping algorithms show better performance than baseline algorithms |
| Alghamdi et al. (2017) |
USA |
32,555 |
Binary |
The proposed ensemble approach achieved high accuracy of prediction (AUC = 0.920) |
| Esteban et al. (2017) |
Argentina |
2,463 |
Multi-class |
The stacked generalization strategy and feed-forward neural network performed the best with validation set |
| Anderson et al. (2015) |
USA |
24,331 |
Binary |
The proposed ensemble model accurately predicted progression to T2DM (AUC = 0.76), and was validated out of sample (AUC = 0.78) |
