Table 2 Classification models performance metrics.

From: Self-report symptom-based endometriosis prediction using machine learning

 

1

Decision Tree

2

Random Forest

3

Gradient Boosting

4

AdaBoost

Recall (sensitivity)

0.890 (0.035)

0.924 (0.029)

0.924 (0.02)

0.939 (0.029)

Specificity

0.859 (0.039)

0.937 (0.031)

0.932 (0.051)

0.934 (0.052)

Precision

0.880 (0.029)

0.945 (0.026)

0.942 (0.042)

0.944 (0.042)

F1-score

0.885 (0.019)

0.934 (0.02)

0.932 (0.021)

0.941 (0.029)

Accuracy

0.876 (0.02)

0.930 (0.022)

0.928 (0.024)

0.937 (0.032)

AUC

0.875 (0.02)

0.930 (0.022)

0.928 (0.025)

0.937 (0.033)

  1. This table shows the predictive performance across four classification models (1) Decision tree, (2) Random Forest, (3) Gradient Boosting, (4) AdaBoost. For each metric we present the mean value and standard deviation based on ten-fold cross-validation.
Back to article page