Table 1 Description of previous studies on LBW prediction.
Reference | Objective | Algorithms used | Outcome |
|---|---|---|---|
Reza, T.B. et al.20 | LBW feature selection and prediction using ML | Boruta algorithm and Wrapper method Several ML classifiers | The wrapper method is the most effective way to select features. RF classification worked best for classification |
Ranjbar A et al.21 | LBW prediction with ML | 8 different learning models | Classification with XGBoost outperformed all others |
de Morais FL.et al.22 | LBW prediction with tree-based ML model | 5 ML classifiers | Attribute selection and eliminating duplicate data improves the model |
Naimi, A.I et al.23 | Fetal growth prediction with ML | Regression-based and data mining techniques | Smoking while pregnant raised the chances of SGA |
Tao, J., et al.24 | Fetal birthweight prediction by a temporal ML method | Convolutional Neuron Networks (CNN), RF, Linear-Regression, Support Vector Regression (SVR), Back Propagation Neural Network (BPNN), and hybrid-LSTM | Hybrid-LSTM has the highest accuracy of 93.3 |
Rubaiya et al.25 | Unravelling birthweight determinants: Integrating ML, spatial analysis, and district-level mapping | Regression tree | Creates maps at the district level for regions that are at high risk of LBW |