Multi-dimensional network data anomaly detection method and device, equipment and storage medium

The invention relates to the artificial intelligence technology, and discloses a multi-dimensional network data anomaly detection method, device and equipment and a medium, and the method comprises the steps: obtaining target historical network data according to a time window, and marking historical anomaly points; performing feature extraction on the target historical network data and sorting according to time to obtain a target historical network data feature sequence, and pre-training an anomaly detection model by using the data to obtain an initial model; fitting abnormal values corresponding to the historical abnormal points to obtain an abnormal point score interval in a preset time period; extracting real-time network data characteristics of the real-time network data, and performing forgetting operation on the initial model according to the real-time network data characteristics to obtain an updated model; calculating a target complexity difference between the update model and the initial model to obt