Multiple Attack Detection Method of Power Intelligent Terminal Based on LSTM Neural Network

Abstract:
Due to the differentiated characteristics of network attacks, the detection effect of multiple attacks is not ideal. Therefore, a multiple attack detection method of power intelligent terminal based on LSTM neural network is proposed. The main problems of over fitting and gradient disappearance in multiple attack detection are analyzed. The one hot coding method is used to numerically process the Bayesian processed data. After min max standardization, matrix decomposition MF is used to find the internal potential characteristics of the data, K-means clustering is carried out, and the data to be identified is input into LMST neural network, Update the parameters until the actual output is consistent with the ideal output to complete the attack detection. Experimental results show that the attack detection rate of this method is as high as 95.29%, and the attack error detection rate is 3.63%, which has a good attack detection effect.