Hybrid Low-Order and Higher-Order Graph Convolutional Networks

Table 1

Text network datasets.

Datasets	C	D	Tr	Te	N
R52	52	9,100	6,532	2,568	17,992
OH	23	7,400	3,357	4,043	21,557
20NG	20	18,846	11,314	7,532	61,603
R8	8	7,674	5,485	2,189	15,362
MR	2	10,662	7,108	3,554	29,426

C indicates the category, D is the total number of texts, Tr is the training set, Te is the test set, and N is the number of vertices of the graph network.