GSA-KAN: A Hybrid Model for Short-Term Traffic Forecasting

Short-term traffic flow forecasting is an essential part of intelligent transportation systems. However, it is challenging to model traffic flow accurately due to its rapid changes over time. The Kolmogorov–Arnold Network (KAN) has shown parameter efficiency with lower memory and computational overh...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhizhe Lin, Dawei Wang, Chuxin Cao, Hai Xie, Teng Zhou, Chunjie Cao
Format: Article
Language:English
Published: MDPI AG 2025-03-01
Series:Mathematics
Subjects:
Online Access:https://www.mdpi.com/2227-7390/13/7/1158
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850212657505763328
author Zhizhe Lin
Dawei Wang
Chuxin Cao
Hai Xie
Teng Zhou
Chunjie Cao
author_facet Zhizhe Lin
Dawei Wang
Chuxin Cao
Hai Xie
Teng Zhou
Chunjie Cao
author_sort Zhizhe Lin
collection DOAJ
description Short-term traffic flow forecasting is an essential part of intelligent transportation systems. However, it is challenging to model traffic flow accurately due to its rapid changes over time. The Kolmogorov–Arnold Network (KAN) has shown parameter efficiency with lower memory and computational overhead via spline-parametrized functions to handle high-dimensional temporal data. In this paper, we propose to unlock the potential of the Kolmogorov–Arnold network for traffic flow forecasting by optimizing its parameters with a heuristic algorithm. The gravitational search algorithm learns to understand optimized KANs for different traffic scenarios. We conduct extensive experiments on four real-world benchmark datasets from Amsterdam, the Netherlands. The RMSE of GSA-KAN is reduced by <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>3.95</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>6.96</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>2.71</mn><mo>%</mo></mrow></semantics></math></inline-formula>, and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>2.29</mn><mo>%</mo></mrow></semantics></math></inline-formula>, and the MAPE of GSA-KAN is reduced by <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>6.66</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.88</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>6.41</mn><mo>%</mo></mrow></semantics></math></inline-formula>, and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>4.87</mn><mo>%</mo></mrow></semantics></math></inline-formula> on the A1, A2, A4, and A8 datasets, respectively. The experimental results demonstrate that GSA-KAN performs advanced parametric and nonparametric models.
format Article
id doaj-art-d4fa0e8db01b4ca5a6325b2fcfec2a4e
institution OA Journals
issn 2227-7390
language English
publishDate 2025-03-01
publisher MDPI AG
record_format Article
series Mathematics
spelling doaj-art-d4fa0e8db01b4ca5a6325b2fcfec2a4e2025-08-20T02:09:17ZengMDPI AGMathematics2227-73902025-03-01137115810.3390/math13071158GSA-KAN: A Hybrid Model for Short-Term Traffic ForecastingZhizhe Lin0Dawei Wang1Chuxin Cao2Hai Xie3Teng Zhou4Chunjie Cao5School of Cyberspace Security, Hainan University, Haikou 570228, ChinaSchool of Cyberspace Security, Hainan University, Haikou 570228, ChinaSchool of Information and Communication Engineering, Hainan University, Haikou 570228, ChinaSchool of Cyberspace Security, Hainan University, Haikou 570228, ChinaSchool of Cyberspace Security, Hainan University, Haikou 570228, ChinaSchool of Cyberspace Security, Hainan University, Haikou 570228, ChinaShort-term traffic flow forecasting is an essential part of intelligent transportation systems. However, it is challenging to model traffic flow accurately due to its rapid changes over time. The Kolmogorov–Arnold Network (KAN) has shown parameter efficiency with lower memory and computational overhead via spline-parametrized functions to handle high-dimensional temporal data. In this paper, we propose to unlock the potential of the Kolmogorov–Arnold network for traffic flow forecasting by optimizing its parameters with a heuristic algorithm. The gravitational search algorithm learns to understand optimized KANs for different traffic scenarios. We conduct extensive experiments on four real-world benchmark datasets from Amsterdam, the Netherlands. The RMSE of GSA-KAN is reduced by <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>3.95</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>6.96</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>2.71</mn><mo>%</mo></mrow></semantics></math></inline-formula>, and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>2.29</mn><mo>%</mo></mrow></semantics></math></inline-formula>, and the MAPE of GSA-KAN is reduced by <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>6.66</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.88</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>6.41</mn><mo>%</mo></mrow></semantics></math></inline-formula>, and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>4.87</mn><mo>%</mo></mrow></semantics></math></inline-formula> on the A1, A2, A4, and A8 datasets, respectively. The experimental results demonstrate that GSA-KAN performs advanced parametric and nonparametric models.https://www.mdpi.com/2227-7390/13/7/1158traffic flow theoryintelligent transportationKolmogorov–Arnold networksgravitational search algorithm
spellingShingle Zhizhe Lin
Dawei Wang
Chuxin Cao
Hai Xie
Teng Zhou
Chunjie Cao
GSA-KAN: A Hybrid Model for Short-Term Traffic Forecasting
Mathematics
traffic flow theory
intelligent transportation
Kolmogorov–Arnold networks
gravitational search algorithm
title GSA-KAN: A Hybrid Model for Short-Term Traffic Forecasting
title_full GSA-KAN: A Hybrid Model for Short-Term Traffic Forecasting
title_fullStr GSA-KAN: A Hybrid Model for Short-Term Traffic Forecasting
title_full_unstemmed GSA-KAN: A Hybrid Model for Short-Term Traffic Forecasting
title_short GSA-KAN: A Hybrid Model for Short-Term Traffic Forecasting
title_sort gsa kan a hybrid model for short term traffic forecasting
topic traffic flow theory
intelligent transportation
Kolmogorov–Arnold networks
gravitational search algorithm
url https://www.mdpi.com/2227-7390/13/7/1158
work_keys_str_mv AT zhizhelin gsakanahybridmodelforshorttermtrafficforecasting
AT daweiwang gsakanahybridmodelforshorttermtrafficforecasting
AT chuxincao gsakanahybridmodelforshorttermtrafficforecasting
AT haixie gsakanahybridmodelforshorttermtrafficforecasting
AT tengzhou gsakanahybridmodelforshorttermtrafficforecasting
AT chunjiecao gsakanahybridmodelforshorttermtrafficforecasting