Skip to main content

Table 12 Correlations of each transferability measure with DeepSleepNet accuracy for individual algorithms as well as overall

From: Comparison of deep transfer learning algorithms and transferability measures for wearable sleep staging

Measure

Head Re-train, \({n = 144}\)

CORAL, \( {n = 144}\)

Per-Class CORAL, \({n = 144}\)

SA, \( {n = 144}\)

DDC, \({n = 144}\)

Overall, \({n = 576}\)

LEEP

0.01

0.13

0.11

0.36***

N/A

0.15****

H-score

0.12

0.15

0.23*

0.26***

N/A

0.19***

Hypothesis margin

− 0.32***

− 0.16*

− 0.27

0.10

N/A

− 0.16***

Silhouette score

− 0.31***

− 0.21*

− 0.21*

0.13

0.07

− 0.11*

MMD, \(\gamma = 0.1\Gamma \)

− 0.40***

− 0.31***

− 0.46***

− 0.18*

N/A

− 0.34***

MMD, \(\gamma = 1\Gamma \)

− 0.28***

− 0.29***

− 0.40***

− 0.29***

N/A

− 0.31***

MMD, \(\gamma = 10\Gamma \)

− 0.12

-0.13

− 0.28***

- 0.13

N/A

− 0.17***

TDAS, \(\epsilon = 0.1m\)

0.48***

0.47***

0.40***

0.30***

N/A

0.40***

TDAS, \(\epsilon = 1m\)

0.38***

0.37***

0.31***

0.25**

N/A

0.33***

TDAS, \(\epsilon = 10m\)

− 0.11

− 0.09

− 0.10

0.07

N/A

− 0.06

  1. *p < 0.05
  2. **p < 0.01
  3. ***p < 0.001