Metalearning-based strategy game behavior prediction method and predictor

The invention relates to a behavior prediction method and predictor of a strategy game based on meta-learning in the technical field of strategy games. The method comprises the steps of obtaining a network attack and defense game data set, dividing the data set into a new task used for testing the effect of a trained behavior prediction model and a training sample used for meta-learning, and training the constructed behavior prediction model based on a deep neural network by adopting a meta-learning method, and the trained behavior prediction model is used to predict the strategic behavior of the attacker in the network attack and defense game in the new task. According to the method, a task classification method of unsupervised learning and a meta-learning method of an expert mixed architecture are adopted, the prediction precision and prediction speed of strategic behaviors of attackers in a network attack and defense game can be obviously improved in a scene with a small network attack and defense game dat