Abnormal flow detection method and device, and medium

The invention relates to an abnormal traffic detection method and device, and a medium. The method comprises the following steps: acquiring network traffic data; a traffic classification result is obtained through a trained abnormal traffic detection model, an attention mechanism and batch standardization are added, and the training method of the abnormal traffic detection model comprises the following steps: performing feature extraction and data preprocessing on a training sample set to obtain normalized numerical features; performing CNN and LSTM on the numerical features in sequence to obtain two-dimensional spatial features and time sequence features, and performing Flatten layer on the two-dimensional spatial features to obtain one-dimensional spatial features; an attention mechanism is added between each full-connection layer and an excitation function in the abnormal traffic detection model according to attention weight enhancement related features and batch standardization; and the traffic trend pred