Identifying and Classifying Enhancers by Dinucleotide-Based Auto-Cross Covariance and Attention-Based Bi-LSTM

Table 3

The performance comparison of different parameters in the word2vec model in the enhancer classification in the 5-fold cross-validation.

	HS/NS	Min_count, Window	ACC (%)	MCC
CBOW	HS	3, 3	65.10	0.3123
		3, 5	66.22	0.3412
		5, 3	65.20	0.3119
		5, 5	67.57	0.3529
	NS	3, 3	61.82	0.2365
		3, 5	66.89	0.3381
		5, 3	63.76	0.2761
		5, 5	63.85	0.2908
Skip-gram	HS	3, 3	66.22	0.3258
		3, 5	65.54	0.3113
		5, 3	66.89	0.3379
		5, 5	65.20	0.3178
	NS	3, 3	66.44	0.3414
		3, 5	63.76	0.2768
		5, 3	64.09	0.2833
		5, 5	63.18	0.2754