Network attack detection method based on deep reinforcement learning

The invention discloses a network attack detection method based on deep reinforcement learning. The method comprises the following steps: preprocessing an original data set, and constructing an Agent, including initializing an environment in which the Agent is located, and specifying an interaction mode, a training strategy and a value function of an intelligent agent and the environment; and selecting the features according to the state, and inputting the selected features into a detection model for prediction. And the detection result is used as feedback to be transmitted back to the agent training module, Q (s, a) of the action is calculated, and a Q table is refreshed. The steps are repeated until the number of features contained in the optimal feature subset reaches the maximum, namely, the model is converged; or the training step length is completed, and an optimal feature subset is generated. The processing method designed for the novel features can reflect the importance of the ...