Table 1.
Summary of PMA datasets used (2019–2022).
Fig 1.
Data preprocessing flowchart.
Fig 2.
Distribution of the output variable before and after SMOTE.
Table 2.
Model configurations and hyperparameters for hyperparameter tuning.
Fig 3.
Methodological work flow.
Table 3.
Socio-demographic characteristics of survey respondents.
Table 4.
Contraceptive knowledge, usage, and exposure to Family Planning (FP) advertising among women.
Table 5.
Model performance of models on default hyperparameters.
Fig 4.
ROC Curve results on Default Hyperparameter.
Fig 5.
Precision-Recall trade-off on Default Hyperparameters.
Table 6.
Models performance after hyperparameter tuning.
Fig 6.
ROC Curve Results after Hyperparameter Tuning.
Fig 7.
Precision-Recall Trade-off after Hyperparameter Tuning.
Table 7.
Performance metrics of classification models after hyperparameter tuning and SMOTEENN balancing.
Fig 8.
Precision-Recall curves for various classification models after hyperparameter tuning and SMOTENN balancing.
Fig 9.
Classification report of the best-performing model (Random Forest).
Fig 10.
Global Feature Importance via SHAP Values.
Fig 11.
SHAP Force Plot for Individual Prediction.