Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network model
In recent decades, traffic flow modelling has become increasingly significant for improving road transportation systems and mitigating congestion on freeways. This research presents a comparative analysis of two machine learning methodologies—Improved Group Method of Data Handling (GMDH) and Artific...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-03-01
|
| Series: | Results in Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2590123025001823 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1825199419223441408 |
|---|---|
| author | Isaac Oyeyemi Olayode Alessandro Severino Frimpong Justice Alex Elmira Jamei |
| author_facet | Isaac Oyeyemi Olayode Alessandro Severino Frimpong Justice Alex Elmira Jamei |
| author_sort | Isaac Oyeyemi Olayode |
| collection | DOAJ |
| description | In recent decades, traffic flow modelling has become increasingly significant for improving road transportation systems and mitigating congestion on freeways. This research presents a comparative analysis of two machine learning methodologies—Improved Group Method of Data Handling (GMDH) and Artificial Neural Network (ANN)—for modelling vehicular traffic flow on a six-lane freeway. The primary objective of this study was to evaluate the predictive accuracy and efficacy of both models in replicating complex traffic patterns and to provide insights into their suitability for real-time traffic flow applications. Traffic flow data were obtained from a six-lane freeway during off-peak and on-peak hours using South African road transportation systems as a case study. Traffic flow variables, such as vehicle density, speed, time, and traffic volume, were considered as both inputs and outputs. The models were trained and validated using this dataset, and the GMDH and ANN were assessed according to their regression efficacy R2and MSE. The results indicate that both models can effectively capture the nonlinear relationships present in the traffic flow of vehicles on a six-lane freeway. However, GMDH outperformed ANN in terms of accuracy and computational efficiency. The optimal regression values for GMDH and ANN were 0.99372 and 0.9167, respectively, demonstrating that GMDH provided a substantially superior fit to the observed data. The exceptional efficacy of the GMDH is attributed to its self-organising architecture and capacity to autonomously identify the most pertinent inputs, thereby reducing model complexity and enhancing generalisation. Artificial Neural Networks, while efficient, require comprehensive tuning and may experience overfitting in high-dimensional datasets. This study suggests that GMDH is a more reliable and effective model for modelling traffic flow on a six-lane freeway, presenting opportunities for real-time traffic prediction and traffic flow management applications. |
| format | Article |
| id | doaj-art-63f31868da634c25baee4d56605a3994 |
| institution | Kabale University |
| issn | 2590-1230 |
| language | English |
| publishDate | 2025-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Results in Engineering |
| spelling | doaj-art-63f31868da634c25baee4d56605a39942025-02-08T05:00:56ZengElsevierResults in Engineering2590-12302025-03-0125104094Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network modelIsaac Oyeyemi Olayode0Alessandro Severino1Frimpong Justice Alex2Elmira Jamei3School of Science, Computing and Engineering Technologies, Swinburne University of Technology, Hawthorn, VIC 3122, Australia; Corresponding author.Department of Civil Engineering and Architecture, University of Kore Di Enna, Kore, ItalySchool of Automotive Engineering, Wuhan University of Technology, Wuhan 430070, ChinaBuilt Environment Discipline, College of Sport, Health and Engineering, Institute for Sustainable Industries and Liveable Cities, Victoria University, Melbourne, VIC 3011, AustraliaIn recent decades, traffic flow modelling has become increasingly significant for improving road transportation systems and mitigating congestion on freeways. This research presents a comparative analysis of two machine learning methodologies—Improved Group Method of Data Handling (GMDH) and Artificial Neural Network (ANN)—for modelling vehicular traffic flow on a six-lane freeway. The primary objective of this study was to evaluate the predictive accuracy and efficacy of both models in replicating complex traffic patterns and to provide insights into their suitability for real-time traffic flow applications. Traffic flow data were obtained from a six-lane freeway during off-peak and on-peak hours using South African road transportation systems as a case study. Traffic flow variables, such as vehicle density, speed, time, and traffic volume, were considered as both inputs and outputs. The models were trained and validated using this dataset, and the GMDH and ANN were assessed according to their regression efficacy R2and MSE. The results indicate that both models can effectively capture the nonlinear relationships present in the traffic flow of vehicles on a six-lane freeway. However, GMDH outperformed ANN in terms of accuracy and computational efficiency. The optimal regression values for GMDH and ANN were 0.99372 and 0.9167, respectively, demonstrating that GMDH provided a substantially superior fit to the observed data. The exceptional efficacy of the GMDH is attributed to its self-organising architecture and capacity to autonomously identify the most pertinent inputs, thereby reducing model complexity and enhancing generalisation. Artificial Neural Networks, while efficient, require comprehensive tuning and may experience overfitting in high-dimensional datasets. This study suggests that GMDH is a more reliable and effective model for modelling traffic flow on a six-lane freeway, presenting opportunities for real-time traffic prediction and traffic flow management applications.http://www.sciencedirect.com/science/article/pii/S2590123025001823Traffic flowGroup method of data handlingMachine learningFreewayArtificial neural network |
| spellingShingle | Isaac Oyeyemi Olayode Alessandro Severino Frimpong Justice Alex Elmira Jamei Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network model Results in Engineering Traffic flow Group method of data handling Machine learning Freeway Artificial neural network |
| title | Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network model |
| title_full | Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network model |
| title_fullStr | Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network model |
| title_full_unstemmed | Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network model |
| title_short | Traffic flow modelling of vehicles on a six lane freeway: Comparative analysis of improved group method of data handling and artificial neural network model |
| title_sort | traffic flow modelling of vehicles on a six lane freeway comparative analysis of improved group method of data handling and artificial neural network model |
| topic | Traffic flow Group method of data handling Machine learning Freeway Artificial neural network |
| url | http://www.sciencedirect.com/science/article/pii/S2590123025001823 |
| work_keys_str_mv | AT isaacoyeyemiolayode trafficflowmodellingofvehiclesonasixlanefreewaycomparativeanalysisofimprovedgroupmethodofdatahandlingandartificialneuralnetworkmodel AT alessandroseverino trafficflowmodellingofvehiclesonasixlanefreewaycomparativeanalysisofimprovedgroupmethodofdatahandlingandartificialneuralnetworkmodel AT frimpongjusticealex trafficflowmodellingofvehiclesonasixlanefreewaycomparativeanalysisofimprovedgroupmethodofdatahandlingandartificialneuralnetworkmodel AT elmirajamei trafficflowmodellingofvehiclesonasixlanefreewaycomparativeanalysisofimprovedgroupmethodofdatahandlingandartificialneuralnetworkmodel |