Traffic flow prediction method based on multi-dimensional time and space dependence mining
The invention relates to a traffic flow prediction method based on multi-dimensional time and space dependency mining, which comprises the following steps of: introducing time-space relationship capture into a traffic flow prediction model, constructing a time dependency extraction sub-model of the...
Gespeichert in:
Hauptverfasser: | , , , , , , , , , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The invention relates to a traffic flow prediction method based on multi-dimensional time and space dependency mining, which comprises the following steps of: introducing time-space relationship capture into a traffic flow prediction model, constructing a time dependency extraction sub-model of the model by utilizing a time convolutional network (TCN) and a sequence attention capture network based on self-attention, and extracting the time dependency extraction sub-model of the model. And optimizing the extraction capability of the model on the time dependence relationship implied in the traffic data. The spatial feature construction process is carried out by utilizing a graph convolutional neural network based on a static graph and a generated graph, modeling analysis is carried out on a dominant spatial relationship and an implicit spatial relationship of a road respectively, and then the dominant spatial relationship and the implicit spatial relationship are fused to generate a complete spatial relationshi |
---|