Traffic flow prediction method and system based on space-time correlation and error compensation

The invention discloses a traffic flow prediction method and system based on space-time correlation and error compensation, and relates to the field of traffic flow prediction, and the method comprises the steps: training a ConvLSTM-based module through employing workday and weekend traffic flow data sets in sequence; the workday traffic flow data set and the weekend traffic flow data set comprise traffic flows at different observation points at different time in a set area; obtaining prediction errors corresponding to the traffic flow sample data; clustering the traffic flow data set by adopting a clustering algorithm; determining a prediction error mean value corresponding to each data category; inputting to-be-predicted data traffic flow into the trained ConvLSTM module to obtain traffic flow prediction data; determining the data category of the to-be-predicted data traffic flow; and performing error compensation on the traffic flow prediction data by adopting the prediction error mean value of the data ca