Fig 1.
Global comparison and gender-specific distribution of overweight and obesity in Bangladesh.
(a) Comparison between overweight/obesity in the World and Bangladesh; and (b) Overweight/obesity by gender in Bangladesh.
Table 1.
BMI was classified according to the WHO.
Table 2.
Comprehensive descriptions of the dataset.
Fig 2.
Feature selection for classification.
Table 3.
Percentages of data distribution before and after applying SMOTE-ENN.
Table 4.
Accuracy (%) of balanced and imbalanced data for the default model.
Table 5.
Optimized hyperparameters after using the SMOTE-ENN technique.
Table 6.
Accuracy (%) of the default and HP-tuned data for the balancing dataset.
Fig 3.
Performance comparison of ML classifiers for the Chi-square (Filter).
Fig 4.
Performance comparison of ML classifiers for the LASSO (Embedded).
Fig 5.
Performance comparison of ML classifiers for the SFS (Wrapper).
Table 7.
Existing studies on overweight and obesity among ever-married women.
Fig 6.
Number of selected features using three feature selection methods.
Fig 7.
Permutation importance of features applied to the proposed SVM model.
Fig 8.
SHAP-based global feature importance and feature-wise contribution patterns for the SVM model.
(a) Global feature importance ranked by mean absolute SHAP values for the SVM model.; (b) SHAP violin plot illustrating the feature-wise contribution patterns and their influence on the model output.