Chat scene abnormal user discovery method and system based on graph neural network

The invention relates to a chat scene abnormal user discovery method and system based on a graph neural network. The method comprises the following steps: respectively processing different given group chat information into a semantic graph and a time sequence graph; training a graph neural network model and a classifier model in the group chat information with the label, wherein the graph neural network model learns user node representations in a semantic graph and a time sequence graph; training a graph neural network model in the group chat information without the label by using an adversarial generation mode, and aligning the graph neural network model with a graph neural network model trained in the group chat information with the label; and classifying the user nodes by using the trained classifier model according to the user node representation learned by the graph neural network model, thereby completing the detection of the abnormal users in all group chat information. According to the method, associa