Logistics demand prediction method and system based on gating circulation unit

The invention provides a logistics demand prediction method and system based on a gating circulation unit, and relates to the field of logistics demand prediction. The method comprises the following steps: acquiring feature data; inputting the feature data into a pre-trained gating circulation unit network model, and obtaining target data output by the pre-trained gating circulation unit network model as a logistics demand result; wherein the process of processing the feature data by the pre-trained gating loop unit network model comprises the step of processing the feature data based on a feature attention mechanism and a time sequence attention mechanism. Based on the above processing, an attention mechanism is introduced, so that the pre-trained gating cycle unit network model focuses on more key information in numerous input feature data, the information overload problem can be solved, and the data processing efficiency and the accuracy of a target data prediction result can be improved. 本发明提供了基于门控循环单元的物流