Table 3 Performance metrics when including the first n important features of each model.

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

 

1

Decision Tree

n = 14

2

Random Forest

n = 55

3

Gradient Boosting

n = 26

4

AdaBoost

n = 24

Recall (sensitivity)

0.893 (0.05)

0.926 (0.037)

0.93 (0.024)

0.932 (0.026)

Specificity

0.903 (0.045)

0.949 (0.018)

0.932 (0.046)

0.946 (0.038)

Precision

0.915 (0.036)

0.955 (0.015)

0.942 (0.036)

0.954 (0.032)

F1-score

0.903 (0.03)

0.94 (0.019)

0.936 (0.019)

0.943 (0.023)

Accuracy

0.897 (0.031)

0.937 (0.019)

0.931 (0.022)

0.939 (0.024)

AUC

0.898 (0.031)

0.938 (0.018)

0.931 (0.023)

0.939 (0.025)

  1. The value of n is indicated in the header of each column. For each metric, we present the mean value and standard deviation based on ten-fold cross-validation.
Back to article page