Intrusion Detection Model for Industrial Internet of Things Based on Improved Autoencoder

Algorithm 1

Training algorithm of intrusion detection model based on SSAE network model.

	Input: 256 dimensional data after high-dimensional mapping and normalization, data with a certain noise proportion .
	Output: optimal network parameter values , , , , and .
	Step 1: the feature extraction model based on SSAE network takes the training data as the input. Through the SGD descent method, the input data are analyzed and processed to obtain the network parameters of the hidden layer. Finally, the output of the first hidden layer is calculated by using the original data and parameters .
	Step 2: then, combined with and , the output parameter and output of the hidden layer can be obtained through the calculation and analysis of the second layer.
	Step 3: repeat step 1 and step 2, and get the weight parameters , , , and by layer-by-layer training. With the help of the calculation and analysis of the classifier, the parameter is obtained.
	Step 4: through the above calculation, we can obtain the network parameter of the detection model. By introducing random noise, we input it as training data, calculate the loss function between the predicted value and the target, and use various optimization methods to calculate the parameters near the minimum value.