Traffic accident occurrence area prediction method and device

The invention discloses a traffic accident occurrence area prediction method and device. The method comprises the steps that a first traffic characteristic graph of a first area in a first time period is acquired, the first traffic characteristic graph is used for representing influence factors influencing traffic accidents in the first area, and the influence factors comprise at least one of the following factors: a vehicle movement track, weather, traffic flow, a road structure and a road image; determining a first feature matrix corresponding to the first traffic feature map; and performing prediction processing on the first feature matrix by using a pre-trained target traffic accident prediction model to obtain a first prediction sub-region in which the traffic accident occurs in the first region. According to the method and the device, the technical problem that the accuracy of a prediction result is relatively poor due to the fact that spatial heterogeneity and temporal heterogeneity among the sub-regio