Table 3.

Performance on classification accuracy of one-dimensional CNNs

Method	Accuracy	F-value
CNN with DAFS (word2vec)	0.980	0.931
CNN with DAFS (one-hot coding)	0.971	0.901
CNN with DAFS (only secondary structure)	0.943	0.803
CNN with Clustal Omega (one-hot coding)	0.958	0.850

Note: ‘CNN with DAFS (word2vec)’ represents one-dimensional CNN with input of DAFS alignments and word2vec distributed representation, ‘CNN with DAFS (one-hot coding)’ denotes the one with input of DAFS alignments and one-hot coding distributed representation, ‘CNN with DAFS (only secondary structure)’ represents the one with input of only secondary structure information, and ‘CNN with Clustal Omega (one-hot coding)’ denotes the one with input of Clustal-Omega alignments and one-hot coding distributed representation.