基于联邦学习的无线通信网络DoS攻击检测方法OA北大核心CSTPCD

A DoS attack on wireless communication network can increase network load,resulting in more delay.In wireless communication network,data is usually scattered across multiple nodes,which can lead to data leakage and attack.Therefore,a method of wireless communication network DoS attack detection based on federation learning is proposed.The initial wireless communication network data is preprocessed and normalized,and the random forest algorithm is used for the dimensionality reduction to remove redundant feature and obtain the optimal network data feature set.The feature set is input into the federation learning training model with deep convolutional neural network as the general model,to independently train the local model and perform the model modification.It is transmitted to the central server for aggregation,and the training is completed after convergence.The trained federation learning model is used to detect the DoS attack rate in wireless communication network,and compared with the maximum capacity received by the receiver to determine whether there is a DoS attack.The experimental results show that the proposed method has high stability and reliability when processing large amounts of data,and can accurately detect DoS attack in a short time.