Table 2 Prediction accuracy achieved with different kernel functions (Unit: ns).

From: Combining semiparametric and machine learning approaches for short-term prediction of satellite clock bias

Clock Type

PRN

Semi-K

kernel1

kernel2

kernel3

MEO Rubidium

C19

0.95

0.66

0.77

C20

0.40

0.87

0.44

C21

0.70

0.57

0.56

C22

1.38

0.67

1.71

C23

1.40

1.19

1.11

C24

0.47

0.64

4.69

C32

0.70

1.00

0.36

C33

1.26

0.69

0.99

C36

0.34

0.35

1.14

C37

1.35

0.96

0.93

C41

0.82

0.78

3.29

C42

1.64

1.23

0.70

Average

0.95

0.80

1.39

MEO Hydrogen

C25

0.55

0.78

0.97

C26

0.12

0.06

0.45

C27

0.42

0.18

0.70

C28

0.30

0.25

0.53

C29

0.33

0.36

0.16

C30

0.55

0.15

0.45

C34

0.50

0.93

0.61

C43

0.45

0.56

1.53

C44

0.33

0.21

0.57

C45

0.24

0.10

0.32

C46

0.46

0.31

0.34

Average

0.38

0.35

0.60

IGSO Hydrogen

C38

0.99

0.83

0.91

C39

0.10

0.42

0.10

C40

0.18

0.28

0.15

Average

0.42

0.51

0.38

GEO Hydrogen

C59

0.57

0.42

0.45

C60

1.19

0.95

0.93

Average

0.88

0.68

0.69

Back to article page