Table 3 Comparison of the discrimination power (AUC) of extracted factors by autoencoder and PCA.

From: Screening for chronic conditions with reproductive factors using a machine learning based approach

 

Autoencoder

PCA

p value

Pathoglycemia

0.827 (0.814–0.838)

0.569 (0.560–0.579)

<0.01

Hypertension

0.809 (0.794–0.821)

0.662 (0.651–0.674)

<0.01

Dyslipidemia

0.801 (0.788–0.813)

0.674 (0.669–0.679)

<0.01

  1. (Note: AUCs with 95% CI are reported, t-test is used to examine the statistical difference).
Back to article page