Skip to main content

Table 6 Model comparison through cross-validation on training data

From: Prediction of contraceptive discontinuation among reproductive-age women in Ethiopia using Ethiopian Demographic and Health Survey 2016 Dataset: A Machine Learning Approach

ML Model

Data

Accuracy score (%)

AUC score

Dummy classifier

Unbalanced

49.9

0.49

Balanced

50.9

0.50

Random Forest

Unbalanced

63.8

0.65

Balanced

67.9*

0.74*

Logistic Regression

Unbalanced

63.5

0.67

Balanced

65.6

0.71

KNN

Unbalanced

59.8

0.60

Balanced

61.9

0.66

SVM

Unbalanced

64.9

0.68

Balanced

58.5

0.70

AdaBoost

Unbalanced

63.5

0.67

Balanced

66.2

0.72

XGBoost

Unbalanced

62.2

0.65

Balanced

67.1

0.72

Artificial Neural Net

(MLPClassifier)

Unbalanced

60.7

0.63

Balanced

65.1

0.71

Naïve Bayes

(GaussianNB)

Unbalanced

59.4

0.61

Balanced

59.1

0.63

  1. *Maximum Performance