Table 5 Clarke error grid analysis for three models using 1-second and 10-second segmentation methods.

From: Non-invasive blood glucose monitoring using PPG signals with various deep learning models and implementation using TinyML

 

1 second segment

10 second segment

Zones (%)

ResNet34

VGG16

CNN-LSTM-ATTENTION

ResNet34

VGG16

CNN-LSTM-ATTENTION

A

72.6

52.3

57.59

68.0

46

53.0

B

25.9

44.4

39.8

31.4

50.9

44

C

0.02

0

0

0.4

0

0

D

1.42

2.5

2.5

0.1

2.4

2.6

E

0.003

0

0

0

0

0

Back to article page