Table 4 Hyper-parameters detail in all classifiers
From: An ensemble deep learning based IDS for IoT using Lambda architecture
Model | Classification | Hyperparameters |
|---|---|---|
ANN | Binary | Four hidden dense layers 128, 256,256 and 128 nodes |
are set with 1 final output dense layer with 1 node | ||
Multi | Four hidden dense layers 128, 256, 256, and 128 nodes | |
are set with 1 final output dense layer with 5 nodes | ||
CNN | Binary | 2 Conv2D filters=32, kernel_size=(1,3) 2 Conv2D with |
filters=64, kernel_size=(1,3) 2 hidden dense layers 256,512 | ||
2 MaxPool2D with 1 final output dense layer with 1 node | ||
Multi | 2 Conv2D filters=32, kernel_size=(1,3) 2 Conv2D with | |
filters=64, kernel_size=(1,3) 2 hidden dense layers 256,512 | ||
2 MaxPool2D with 1 final output dense layer with 5 node | ||
LSTM | Binary | 3 LSTM layers 60,120,120 with final |
output dense layer with 1 node | ||
Multi | 3 LSTM layers 60,120,120 with final | |
output dense layer with 5 node |