Reinforcement learning knowledge graph reasoning method based on generative adversarial imitation learning

The invention provides a knowledge graph reasoning algorithm based on RLKGR-GAIL in combination with a generative adversarial imitation learning method, and the knowledge graph reasoning algorithm is used for improving an existing knowledge graph path reasoning model based on reinforcement learning. The method mainly comprises the steps that an RLKGR-GAIL algorithm is put forward, a corresponding overall framework diagram is given, then a knowledge graph data set in the aircraft maintenance field is constructed according to existing maintenance texts, a contrast experiment is set according to the knowledge graph data set and an existing knowledge graph reasoning algorithm, and the knowledge graph data set is obtained. Experimental results show that the RLKGR-GAIL model has certain advantages in relation link prediction tasks compared with other models, and the superiority of the RLKGR-GAIL model is proved. 本发明结合生成式对抗模仿学习的方法提出一种基于RLKGR-GAIL的知识图谱推理算法用于改进现有的基于强化学习的知识图谱路径推理模型。内容主要包括：提出了RLKGR-GAIL算法，并给出了相应的总体框架图，然