Machine learning to predict unintended pregnancy among reproductive-age women in Ethiopia: evidence from EDHS 2016

Table 3 Compares imbalanced data handling techniques using accuracy and Area under the curve (AUC)

Algorithms	Comparison method	Unbalanced	SMOTE
Logistic Regression	Accuracy (%)	80.25	70.00
Logistic Regression	AUC	0.668	0.775
Decision Tree	Accuracy (%)	66.75	75.95
Decision Tree	AUC	0.557	0.760
Random Forest	Accuracy (%)	79.41	84.40
Random Forest	AUC	0.659	0.924
Gradient Boosting	Accuracy (%)	79.13	74.91
Gradient Boosting	AUC	0.682	0.824
XGBoost	Accuracy (%)	77.32	82.22
XGBoost	AUC	0.641	0.898
Extra Tree classifier	Accuracy (%)	78.74	84.93
Extra Tree classifier	AUC	0.628	0.926

SMOTE: Synthetic Minority Over-sampling Technique, AUC: Area Under Curve, Underline and bold numbers were the highest score of the classifier

ISSN: 1472-6874