Table 12.
Comparative results of ANN with previous studies based on accuracy%.
| Author | Approach | Algorithm | Accuracy% |
|---|---|---|---|
| Alam et al. [5] | Prediction of diabetes | ANN | 75.7% |
| Median values and NB imputation | |||
| Pradhan et al. [61] | Prediction of diabetes with classifier comparisons | ANN | 85.09 |
| Guldogan et al. [62] | Prediction of diabetes: | MLP | 78.1% |
| Missing values deleted | RBF | 76.8% | |
| Ahuja et al. [15] | Prediction of diabetes: | MLP | 78.7% |
| Missing values by median | |||
| Ramezani et al. [63] | Predicting diabetes with reducing number of features from 8 to 5 | LANFIS | 88.05% |
| Our work (ANN) | Predicting diabetes with preprocessing steps + SMOTE 100% | ANN | 90.2% |