Skip to main content
. 2022 Apr 7;2022:9879610. doi: 10.34133/2022/9879610

Figure 2.

Figure 2

ANN with three hidden layers. The letter “I” stands for the input layer, I1 to I38 are the 38 input predictors: H for hidden layers. H1 to H15 are the 15 neurons of the first hidden layer, 10 neurons in the second hidden layer, and five neurons in the third hidden layer. B1, B2 B3, and B4 for bias to indicate weighted connections between layers. O for output layer, O1 to O3 are the three possible predictions for RSA B, T, and TB root types. The final prediction for the sample in the figure is T, a taproot RSA with 97% probability.