A traffic time sequence prediction method based on a gating network and gradient lifting regression

Abstract

The invention discloses a method for predicting a traffic time sequence based on a gating network GRU and a gradient lifting regression model GBR. The method comprises a multi-time dimension data extraction process, a mode mining process, a time sequence data prediction process and a rolling prediction process. Through the multi-time-dimension data extraction, the short-term and long-term mode mining is carried out on time sequence data by using GRU, the preliminary prediction is carried out by using GBR in combination with traffic trend and related road data, and then the preliminary prediction result is fused to obtain a final traffic time sequence data prediction value. According to the method, the potential long-term and short-term modes of the time sequence can be explored, the fine tuning is carried out according to the real-time data, and the method can adapt to the current traffic condition and is high in prediction precision and expandability of the traffic time sequence and has interpretability.

Description

technical field [0001] The invention belongs to the technical field of intelligent transportation, and relates to a traffic time series prediction method, in particular to a traffic time series prediction method based on a gated neural network and gradient boosting regression. Background technique [0002] With the rapid development of the economy, people's living standards have gradually improved. More and more private cars have put a lot of pressure on the traffic network, and traffic congestion can be seen everywhere. The emergence of intelligent transportation system has relieved the traffic pressure and provided convenience for people's life. Traffic forecasting is an essential part of intelligent transportation systems. Traffic time series forecasting generally uses pattern mining for historical and real-time data to predict traffic data for a period of time in the future. [0003] The existing traffic time series data prediction methods mainly include linear regres...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a new traffic time-series data prediction method, based on the data extraction of multiple time dimensions of traffic data to combine prior traffic knowledge, using gated neural network and gradient boosting regression to predict traffic time series It is used to solve the shortcomings of the existing intelligent transportation using neural networks that cannot combine human prior knowledge and surrounding relevant road data for traffic time series prediction, and more accurately predict future traffic time series data (such as travel time, speed and flow, etc.)

Images