Event entity prediction method based on context enhanced tense knowledge graph

The invention discloses an event entity prediction method based on a context-enhanced tense knowledge graph. The event entity prediction method comprises the following steps: constructing a tense knowledge graph and an entity static attribute graph; learning event entity features and event relationship features; historical event frequency information is transmitted; and predicting an event entity according to the event entity feature, the event relationship feature and the historical information frequency. According to the method, the static attribute graph of the event entity is introduced to serve as context information of the event, feature representation of the event entity is supplemented, and then autoregression modeling of the historical knowledge graph is completed in combination with the relation perception GCN and the time gating unit, so that the problem of context information omission in the social network public opinion event entity prediction process is solved, and the prediction efficiency is i