Fig. 3
From: Deep learning models for bacteria taxonomic classification of metagenomic data
The convolutional neural network. The architecture of the convolutional neural network used. Here, L represents the dimension of the input vector x, L=4K were K is the dimension of the K-mers. In the upper part of the figure the representation of the C1 convolutional-maxpooling layer, where K stands for kernel size and n1 is the number of kernels. The block M1 represents the set of weights for the connections from input to hidden layer, the block M2 represents the weighted connections from hidden layer to output. y is the CNN output