Skip to main content
. 2018 May 10;6:86. doi: 10.1186/s40168-018-0469-5

Table 1.

ANN training parameters

Input nodes 62
Hidden 1 nodes 59
Activation 1 Hyperbolic tangent
Hidden 2 nodes 53
Activation 2 Sigmoid
Output nodes 56
Error threshold 10−5
Momentum 0.65
Learning rate 0.15
Max epochs 20,000