Skip to main content

Table 2 Model configurations for MLP and CNN

From: MetaNN: accurate classification of host phenotypes from metagenomic data using neural networks

 

Synthetic

CBH

CSS

HMP

CS

FS

FSH

IBD

PDX

MLP

(256, 256)

(1024, 512)

(512, 256)

(512, 256)

(512, 512)

(512, 512)

(512, 256)

(512, 256, 128)

(512, 256, 128)

CNN

Conv1D(8, 3) → Dropout → ReLu → MaxPool1D(2) → Conv1D(8, 3) → ReLu → MaxPool1D(2) → FC

  1. Number in the round bracket represents the number of hidden units. Conv1D is the one-dimensional convolution layer. ReLu is the non-linear rectifier layer. MaxPool1D represents the one-dimensional max pooling layer. Dropout and FC represent dropout and fully connected layers, respectively. Details of each dataset are described in Table 1
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us