Table 7 Performance of different methods versus proposed GCN-ARE for Botswana Dataset.

From: Collaborative representation and confidence-driven semi-supervised learning for hyperspectral image classification

class

GAT21

ViT25

Hybrid23

GEN19

SCCAN10

EfficientFormer26

GCN-ARE

1

100.00

98.81

97.66

100.00

97.28

96.15

100.00

2

83.51

94.19

92.05

90.00

83.51

92.05

100.00

3

97.88

100.00

99.57

100.00

94.29

100.00

100.00

4

100.00

100.00

100.00

97.94

100.00

100.00

100.00

5

99.52

97.90

98.74

98.08

98.10

98.21

97.40

6

89.62

95.60

97.55

88.00

93.02

94.09

98.88

7

100.00

97.93

97.49

97.92

99.57

98.31

100.00

8

97.86

94.82

98.92

87.56

98.92

97.86

100.00

9

95.11

98.33

97.65

98.99

98.29

97.64

100.00

10

97.29

96.54

97.78

90.83

98.22

94.42

100.00

11

95.16

99.26

97.20

98.84

97.55

97.54

92.13

12

100.00

100.00

100.00

100.00

94.15

100.00

99.45

13

99.20

100.00

97.25

98.41

99.60

98.80

100.00

14

100.00

100.00

100.00

100.00

100.00

96.15

97.89

OA

96.97

98.18

98.05

96.19

97.00

97.41

98.86

AA

97.39

98.45

98.21

96.27

97.33

97.72

98.98

Kappa

0.97

0.98

0.98

0.96

0.97

0.97

0.99