Skip to main content

Table 3 The Clarke’s EGA for Output of NNARX and RNN (Allam et al. 2011a) at Different prdeiction horizons

From: Using nonlinear auto-regressive with exogenous input neural network (NNARX) in blood glucose prediction

PH

% of samples in zone A

% of samples in zone B

% of samples in zone C

% of samples in zone D

%of samples in zone E

NNARX

RNN

NNARX

RNN

NARX

RNN

NARX

RNN

NARX

RNN

15

83.94

100

16.05

0

0

0

0

0

0

0

30

88.5

98.7

11.47

1.3

0

0

0

0

0

0

45

90.8

91.6

9.2

8.4

0

0

0

0

0

0

60

90.02

78.8

9.98

19.3

0

0

0

1.9

0

0

90

85.46

11.71

0

0

0

100

83.73

13.44

0

0.43

0