Research Article
An Efficient Communication Intrusion Detection Scheme in AMI Combining Feature Dimensionality Reduction and Improved LSTM
Figure 11
The running time of each program when the set threshold is reached. (i) The accuracy of the training dataset reaches 0.95. (ii) The accuracy of the testing dataset reaches 0.93. (iii) The loss of the training dataset reaches 0.12. (iv) The loss of the testing dataset reaches 0.15.