Short-Term Traffic Flow Prediction Based on the Optimization Study of Initial Weights of the Attention Mechanism

Tianhe Lan, Xiaojing Zhang (), Dayi Qu, Yufeng Yang and Yicheng Chen
Additional contact information
Tianhe Lan: School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266520, China
Xiaojing Zhang: Journal Editorial Department, Qingdao University of Technology, Qingdao 266520, China
Dayi Qu: School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266520, China
Yufeng Yang: School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266520, China
Yicheng Chen: School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266520, China

Sustainability, 2023, vol. 15, issue 2, 1-16

Abstract: Traffic-flow prediction plays an important role in the construction of intelligent transportation systems (ITS). So, in order to improve the accuracy of short-term traffic flow prediction, a prediction model (GWO-attention-LSTM) based on the combination of optimized attention mechanism and long short-term memory (LSTM) is proposed. The model is based on LSTM and uses the attention mechanism to assign individual weight to the feature information extracted via LSTM. This can increase the prediction model’s focus on important information. The initial weight parameters of the attention mechanism are also optimized using the grey wolf optimizer (GWO). By simulating the hunting process of grey wolves, the GWO algorithm calculates the hunting position of the grey wolf and maps it to the initial weight parameters of the attention mechanism. In this way, the short-time traffic flow prediction model is constructed. The traffic flow data of the trunk roads in the center of Qingdao (China) are used as the research object. Multiple sets of comparison models are set up for prediction analysis. The results show that the GWO-attention-LSTM model has obvious advantages over other models. The prediction error MAE values of the GWO-attention-LSTM model decreased by 7.32% and 14.35% on average compared with the attention-LSTM model and LSTM model. It is concluded that the GWO-attention-LSTM model has better model performance and can provide effective help for traffic management control and traffic flow theory research.

Keywords: intelligent transportation; short-term traffic flow prediction; attention mechanism; long short-term memory; grey wolf optimizer; deep learning (search for similar items in EconPapers)
JEL-codes: O13 Q Q0 Q2 Q3 Q5 Q56 (search for similar items in EconPapers)
Date: 2023
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2071-1050/15/2/1374/pdf (application/pdf)
https://www.mdpi.com/2071-1050/15/2/1374/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jsusta:v:15:y:2023:i:2:p:1374-:d:1032050

Access Statistics for this article

Sustainability is currently edited by Ms. Alexandra Wu

More articles in Sustainability from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().