Localized empirical threat analysis method and device based on neural network

The invention provides a localized experience threat analysis method and device based on a neural network, belongs to the technical field of network security, and particularly relates to network threat analysis. The problems that an existing threat analysis method is low in analysis precision, poor in robustness, low in efficiency, difficult to effectively deal with complex and deep threat analysis and narrow in application range are solved. The method comprises the following steps: S1, preprocessing original network data; s2, performing feature extraction on the to-be-analyzed network data to obtain a feature sequence; s3, generating a malicious behavior sequence according to the feature sequence; and S4, performing analysis according to the malicious behavior sequence by adopting a threat analysis network model to obtain a threat analysis result. The local experience threat analysis method and device based on the neural network are mainly used in the field of network security. 本发明提出了一种基于神经网络的本地化经验威胁分析方法及装置，