EnClaSC: a novel ensemble approach for accurate and robust cell-type classification of single-cell transcriptomes

BMC Bioinformatics

Table 2 The architecture of the neural network

Layers	Setting
Dense_1	{(feature numbers,128), ‘relu’}
Dropout_1	0.25
Dense_2	{(128,64), ‘relu’}
Dropout_2	0.5
Dense_3	{(64,32), ‘relu’}
Dense_4	{(32, number of classes in training sets), ‘softmax’}

ISSN: 1471-2105