Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.

Skip to main content

Table 5 Utility metric results

From: Comparative assessment of synthetic time series generation approaches in healthcare: leveraging patient metadata for accurate data synthesis

Generation Method

Modela

TSTR metric

A1

DGANT

0.802 ± 0.033

WGAN-GPT

0.880 ± 0.002

DGANM

0.931 ± 0.064

WGAN-GPM

0.909 ± 0.065

DGANP

0.983 ± 0.005

WGAN-GPP

0.986 ± 0.002

A2

DGANT

0.900 ± 0.589

WGAN-GPT

0.837 ± 0.109

DGANM

0.869 ± 0.078

WGAN-GPM

0.924 ± 0.049

DGANP

0.906 ± 0.083

WGAN-GPP

0.919 ± 0.000

A3

DGANT

0.919 ± 0.167

WGAN-GPT

0.900 ± 0.156

DGANM

0.893 ± 0.079

WGAN-GPM

0.758 ± 0.125

DGANP

0.957 ± 0.016

WGAN-GPP

0.778 ± 0.120

  1. Values in bold highlight the best-performing approach per trained model and metric
  2. TSTR Train on Synthetic and Test on Real
  3. aThe subscripts on the model column refer to the dataset used for the training step