Regional flow prediction-oriented spatio-temporal global semantic representation learning method
A technology of traffic prediction and semantic representation, applied in the field of regional traffic prediction, it can solve the problems of ignoring correlation, losing time dependency, etc., to achieve the effect of enhancing accuracy
Pending Publication Date: 2022-05-13
CHONGQING UNIV
View PDF0 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, since this method only considers the individual influence of time dependencies of different scales on the prediction target, and ignores the correlation between time dependencies of different scales, it will lose part of the global time dependence
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0087] The present invention will be described in further detail below.
[0088] Existing methods have achieved some success in integrating spatio-temporal information, but the existing models lack sufficient consideration of global information and location information in the time dimension. This problem can be summarized in the following three aspects: a) The model does not consider time The relative position information on the axis leads to the fact that the position features in the flow diagram are not effectively learned. b) It ignores the correlation between temporal dependencies at different scales, resulting in inaccurate representation of global information. c) These models predict the flow graph at the end of the time series, but do not predict more flow graphs before the end of the time series, resulting in ignoring some temporal features during the learning process.
[0089] Based on the above discussion, the present invention proposes a spatio-temporal global sema...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention relates to a spatial-temporal global semantic representation learning method for regional traffic prediction, which comprises the following steps: S1, establishing a spatial-temporal global semantic representation learning model ST-GSP for regional traffic prediction, the ST-GSP comprising a semantic flow encoder, a transformer encoder and a fusion process which are arranged in sequence; the semantic flow encoder encodes the spatial dependency relationship of different distances and the influence of external factors; the transformer encoder is used for capturing the correlation between the time dependency relationships of different scales; the fusion process fuses the external factors on the historical representation and the future time interval so as to obtain the final representation; s2, the ST-GSP is trained by adopting a self-supervised learning method; and S3, inputting historical data before the to-be-predicted time point into the trained ST-GSP, wherein the output of the ST-GSP is the flow of the to-be-predicted time point. According to the method, more detailed time information is used as position codes, and the accuracy of regional flow prediction is enhanced.
Description
technical field [0001] The invention relates to the technical field of regional traffic forecasting, in particular to a spatio-temporal global semantic representation learning method for regional traffic forecasting. Background technique [0002] Regional flow forecasting has great application potential in intelligent traffic management, travel optimization, and public safety. For example, when a hotspot event occurs, government departments can obtain the evolution of regional traffic through regional traffic forecasting, and carry out traffic diversion in advance to prevent dangerous accidents such as fatal stampedes. And online car-hailing platforms (such as Uber and Didi) can also plan the driving route of online car-hailing vehicles in advance through regional traffic forecasts, so as to increase the number of orders received by drivers. The waste home appliance recycling platform can also use regional traffic forecasts to arrange qualified recycling personnel in advanc...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06Q10/04G06Q50/30G06N3/04G06N3/08
CPCG06Q10/04G06N3/08G06N3/048G06N3/045G06Q50/40
Inventor 高旻赵亮王宗威郭林昕周魏熊庆宇赵泉午
Owner CHONGQING UNIV