Skip to main content

Table 5 Accuracy (in terms of RMSE) of the final RNN model with stacked LSTM layers

From: Stacked LSTM based deep recurrent neural network with kalman smoothing for blood glucose prediction

Patient ID

PH = 30 min

PH = 60 min

Smoothed CGM

Raw CGM

Smoothed CGM

Raw CGM

# 559

4.73

17.85

16.17

31.55

# 563

5.74

18.65

16.31

30.42

# 570

4.81

15.94

14.22

25.74

# 575

8.45

20.93

22.12

31.97

# 588

5.10

17.71

15.73

30.45

# 591

6.53

20.35

18.88

31.80

Mean (RMSE)

5.89

18.57

17.24

30.32

  1. The prediction result is for the dataset with raw CGM readings, and Kalman smoothed CGM values, respectively, over 30 min and 60 min of the prediction horizon
Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us