Road intersection congestion prediction method based on time point process neural network model

The invention discloses a road intersection congestion prediction method based on a time point process neural network model. The method comprises the following steps: firstly, in a spatial correlation modeling process, constructing a spatial correlation module by fusing spatial region congestion change modes of a plurality of intersections to a single intersection level; secondly, in the dual-granularity time correlation modeling process, the time granularity of a congestion event is captured through a time point process and further integrated with a gating circulation network unit, a new neural point process gating circulation unit is constructed, modeling is conducted on congestion under different time granularities through the new neural point process gating circulation unit, and the time granularities of the congestion event are obtained; a dual-granularity time correlation module is obtained; and finally, based on the sequence, obtaining a sequence architecture and a space-time correlation module, establ