Database system anomaly detection method and device based on reconstruction adversarial training

The invention discloses a database system anomaly detection method and device based on reconstruction adversarial training, and belongs to the technical field of database system detection.The method comprises the steps that monitoring data of a database system is organized into two-dimensional image modal representation from one-dimensional representation so as to conduct feature extraction, and a training sample set is constructed; the training sample can be learned by adopting a neural network. The training sample set is utilized to train an anomaly detection model based on a reconstruction adversarial training mode, so that the anomaly detection model fully learns rules in input features, the trained anomaly detection model with reconstruction capability and discrimination capability and a dynamic threshold are utilized to complete anomaly detection work, the anomaly detection accuracy is higher, and the speed is higher; in addition, the abnormal threshold is dynamically adjusted, the tedious manual thresh