A smart decision-based population movement prediction method, an apparatus, and a computer device. The method comprises acquiring an urban map (S201); dividing the urban map with urban regions in the urban map as nodes to generate an urban node network (S202); acquiring historical population information of each node in the urban node network (S203); by means of a map neural network, performing computations on the historical population information of each node to obtain spatial features and time sequence features of each node (S204); on the basis of the spatial features and the time sequence features, generating point embedded vectors for each node (S205); and on the basis of the point embedded vectors, generating population movement information, the population movement information serving as node connection lines connecting each node (S206). The present method improves the accuracy of population movement prediction.
展开▼