Table 6 Macro-average F1 of different hidden layers (%).

From: A hybrid re-fusion model for text classification

the dimension of GCNII hidden layer

MR

Ohsumed

R52

R8

128

87.62

62.37

85.38

95.06

200

88.50

63.31

86.41

95.80

256

86.95

61.93

84.94

94.83

Back to article page