An Efficient Communication Intrusion Detection Scheme in AMI Combining Feature Dimensionality Reduction and Improved LSTM

<div>The structure of SAE. The output of the Autoencoder of the previous layer is used as the input of the Autoencoder of the next layer.</div>

Security and Communication Networks

Figure 6: An Efficient Communication Intrusion Detection Scheme in AMI Combining Feature Dimensionality Reduction and Improved LSTM 