Adversarial training method and device for user behavior log anomaly detection model

The invention discloses an adversarial training method and device for a user behavior log anomaly detection model. The adversarial training method comprises the steps of obtaining a user behavior log data stream; based on a preset coding rule, converting the user behavior log data stream into a sample data stream represented by hexadecimal, binding every two adjacent hexadecimal numbers in the sample data stream into a combination code, and then converting the combination code into an index value so as to obtain an index sequence; converting the index sequence into a feature vector sequence based on a pre-training model; and taking the feature vector sequence as the input of a generative adversarial network, and carrying out mutual game by utilizing a generator and a discriminator of the generative adversarial network, thereby carrying out adversarial training on the pre-training model and the generative adversarial network, and taking the finally trained generator of the generative adversarial network as the