Explainable Boosting Machine-Based Analysis of Delay at Signalized Intersections

Öz

Urban traffic signal control plays an important role in improving operational efficiency and reducing unnecessary delays in transportation networks. However, capturing the complex nonlinear relationships between traffic demand and signal timing parameters remains a major challenge under varying operating conditions. This study proposes an explainable artificial intelligence (XAI) framework for traffic signal delay analysis using scenario-driven operational data from a three-leg signalized intersection. An Explainable Boosting Machine (EBM) was developed to model the relationships between cycle time, phase split, arrival flow, and intersection delay, and was comparatively evaluated against Random Forest (RF) and XGBoost models using 5-fold cross-validation. The EBM model achieved high predictive performance (R2 = 0.89, RMSE = 1.94), while simultaneously providing transparent interpretation of variable effects through global importance analysis, shape functions, and pairwise interactions. The results indicate that cycle time and phase allocation are among the key factors influencing delay formation, with nonlinear interaction patterns observed under different traffic demand conditions. An interaction was observed between cycle length and Phase C allocation (north approach turning phase), suggesting a condition-dependent relationship with delay. Comparative analyses further demonstrated a general consistency between EBM and benchmark ensemble models in identifying influential operational variables. The proposed framework provides predictive capability and interpretable insights that may support data-driven traffic signal analysis and operational decision-making.

Anahtar Kelimeler

• signalized intersection
• control delay
• explainable boosting machine
• data-driven modeling
• predictive modeling

Destekleyen Kurum

There is no conflict of interest with any person / institution in the article prepared.

Etik Beyan

There is no need to obtain permission from the ethics committee for the article prepared.

Teşekkür

This article does not include an acknowledgements section.

Sinyalize Kavşaklarda Gecikmenin Açıklanabilir Tabanlı (EBM) Yöntem ile Analizi

Öz

Kentsel trafik sinyal sitemlerinin kontrolü, sistem verimliliğinin artırılmasında ve gereksiz gecikmelerin azaltılmasında önemli bir rol oynamaktadır. Trafik koşullarının giderek daha değişken hale gelmesiyle birlikte, talep ile sinyal zamanlama parametreleri arasındaki karmaşık etkileşimlerin yakalanması da zorlaşmaktadır. Bu çalışma, senaryo temelli işletim verilerine dayanan, trafik analizine yönelik açıklanabilir yapay zekâ (XAI) tabanlı bir çerçeve sunmaktadır. Explainable Boosting Machine (EBM) kullanılarak geliştirilen önerilen model; çevrim süresi, yeşil faz , geliş akımları, doygunluk derecesi ve kavşak gecikmesi arasındaki ilişkileri ortaya koymaktadır. Tahmin doğruluğunun yanı sıra, model özellik önemi ve şekil fonksiyonları aracılığıyla her bir parametrenin gecikme oluşumuna nasıl katkı sağladığını açıklamaktadır. Elde edilen bulgular, açıklanabilir yapay zekâ modellerinin yüksek tahmin doğruluğunu yorumlanabilir biçimde sunabildiğini ve veriye dayalı sinyal kontrol stratejilerinin geliştirilmesini destekleyerek etkin ve sürdürülebilir kentsel hareketliliğe katkı sağladığını göstermektedir.

Destekleyen Kurum

Hazırlanan makalede herhangi bir kişi/kurum ile çıkar çatışması bulunmamaktadır.

Etik Beyan

Hazırlanan makalede etik kurul izni alınmasına gerek yoktur.

Teşekkür

Hazırlanan makalede herhangi bir teşekkür kısmı bulunmamaktadır.

Kaynakça

1. [1] Z. Cakici and Y. S. Murat, "A differential evolution algorithm‐based traffic control model for signalized intersections," Advances in Civil Engineering, vol. 2019, no. 7360939, pp. 1-16, 2019, doi: 10.1155/2019/7360939.
2. [2] A. I. M. Almadi, R. E. Al Mamlook, Y. Almarhabi, I. Ullah, A. Jamal, and N. Bandara, "A Fuzzy-Logic Approach Based on Driver Decision-Making Behavior Modeling and Simulation," Sustainability, vol. 14, no. 14, p. 8874, 2022, doi: 10.3390/su14148874.
3. [3] R. Carillo, F. Cerasuolo, G. Bovenzi, D. Ciuonzo, and A. Pescapè, "Explainable federated class incremental learning for Encrypted Network Traffic classification," Computer Networks, vol. 269, p. 111448, 2025, doi: 10.1016/j.comnet.2025.111448.
4. [4] A. P. Akgüngör and E. Korkmaz, "Analysis and Modelling of the Relationship between Stopped and Control Delays by Differential Evolution Algorithm," The Open Civil Engineering Journal, vol. 10, no. 1, pp. 266-279, 2016, doi: 10.2174/1874149501610010266.
5. [5] HCM, "Highway Capacity Manual," Washington, D.C. :Transportation Research Board, 2000.
6. [6] F. V. Webster, "Traffic signal settings, road research technical paper No: 39," Road Research Laboratory, London, 1958.
7. [7] Akçelik, R., "Traffic Signals: Capacity and Time Analysis, Australian Road Research Board, Research Report ARR No.123," 1981.
8. [8] S. P. Çalışkanelli, F. Atasever Coşkun, and S. Tanyel, "Start-up Lost Time and its Effect on Signalized Intersections in Turkey," Traffic&Transportation, vol. 29, no. 3, pp. 321-329, 2017.

9. [9] M. Çetin and Y. Ş. Murat, "Doygun Akımın Belirlenmesinde Sürücü Davranışına Bağlı Bir Matematik Yaklaşım," İMO Teknik Dergi, vol. 24, no. 117, pp. 6399-6414, 2013.
10. [10] Y. Ş. Murat and Ö. Baskan, "Modeling vehicle delays at signalized junctions: Artificial neural networks approach," Journal of Scientific & Industrial Research, vol. 65, pp. 558-564, 2006.
11. [11] E. Doğan, A. P. Akgüngör, and T. Arslan, "Estimation of delay and vehicle stops at signalized intersections using artificial neural network," Engineering Review, vol. 36, no. 2, pp. 157-165, 2016.
12. [12] P. Ranpura, V. Shukla, and R. Gujar, "Estimation of vehicle control delay using artificial intelligence techniques for heterogeneous traffic conditions," Expert Systems with Applications, vol. 246, p. 123206, 2024, doi: 10.1016/j.eswa.2024.123206.
13. [13] M. E. C. Bagdatli and A. S. Dokuz, "Vehicle Delay Estimation at Signalized Intersections Using Machine Learning Algorithms," Transportation Research Record: Journal of the Transportation Research Board, vol. 2675, no. 9, pp. 110-126, 2021, doi: 10.1177/03611981211036874.
14. [14] Y. Ş. Murat, "Comparison of fuzzy logic and artificial neural networks approaches in vehicle delay modeling," Transportation Research Part C: Emerging Technologies, vol. 14, no. 5, pp. 316-334, 2006, doi: 10.1016/j.trc.2006.08.003.
15. [15] A. Mahmoudian, M. Bypour, and M. Kioumarsi, "Explainable Boosting Machine Learning for Predicting Bond Strength of FRP Rebars in Ultra High-Performance Concrete," Computation, vol. 12, no. 202, pp. 1-35, 2024, doi: 10.3390/computation12100202.
16. [16] H. Wang, Y. Ding, Y. Kong, D. Sun, Y. Shi, and X. Cai, "Predicting the Compressive Strength of Sustainable Portland Cement–Fly Ash Mortar Using Explainable Boosting Machine Learning Techniques," Materials, vol. 17, no. 4744, pp. 1-21, 2024, doi: 10.3390/ma17194744.
17. [17] Z. Liu et al., "Nonlinear Cooling Effect of Street Green Space Morphology: Evidence from a Gradient Boosting Decision Tree and Explainable Machine Learning Approach," Land, vol. 11, no. 2220, pp. 1-23, 2022, doi: 10.3390/land11122220.
18. [18] I. Ahmed, I. Kumara, V. Reshadat, A. S. M. Kayes, W.-J. van den Heuvel, and D. A. Tamburri, "Travel Time Prediction and Explanation with Spatio-Temporal Features: A Comparative Study," Electronics, vol. 11, no. 1, p. 106, 2021, doi: 10.3390/electronics11010106.
19. [19] X. Zhang, "Traffic Flow Prediction Based on Explainable Machine Learning," Highlights in Science, Engineering and Technology, vol. 56, pp. 56-64, 2023, doi: 10.54097/hset.v56i.9816.
20. [20] R. Alahmadi, H. Almujibah, S. Alotaibi, A. E. A. Elshekh, M. Alsharif, and M. Bakri, "Explainable Boosting Machine: A Contemporary Glass-Box Model to Analyze Work Zone-Related Road Traffic Crashes," Safety, vol. 9, no. 4, p. 83, 2023, doi: 10.3390/safety9040083.
21. [21] W.-L. Liu, J. Zhong, P. Liang, J. Guo, H. Zhao, and J. Zhang, "Towards explainable traffic signal control for urban networks through genetic programming," Swarm and Evolutionary Computation, vol. 88, no. 101588, 2024, doi: 10.1016/j.swevo.2024.101588.
22. [22] P. Santhiya et al., "Explainable artificial intelligence for traffic signal detection using LIME algorithm," International Journal of Informatics and Communication Technology (IJ-ICT), vol. 13, no. 3, p. 527, 2024, doi: 10.11591/ijict.v13i3.pp527-536.
23. [23] K. M. Erten and R. Gürfidan, "Regression-Based Performance Prediction in Asphalt Mixture Design and Input Analysis with SHAP," Applied Sciences, vol. 15, no. 19, p. 10779, 2025, doi: 10.3390/app151910779.
24. [24] C. O. Retzlaff et al., "Post-hoc vs ante-hoc explanations: xAI design guidelines for data scientists," Cognitive Systems Research, vol. 86, p. 101243, 2024, doi: 10.1016/j.cogsys.2024.101243.
25. [25] D. Erol and Ö. Başkan, "Developing a Capacity Model for Roundabouts Using SIDRA Calibrated via Simulation-Based Optimization," Sustainability, vol. 17, no. 10289, 2025, doi: 10.3390/su172210289.
26. [26] P. Çalişkanelli and S. Tanyel, "Investigating the driver’s response time at signalized intersections," Transport, vol. 33, no. 2, pp. 380-388, 2016, doi: 10.3846/16484142.2016.1250106.
27. [27] C. Ng, S. Susilawati, and I. C. Mei Leng, "Piecewise Linear Regression: A Statistical Method for the Analysis of the Relationship between Traffic Signal Parameters and Air Pollutant Emissions," in Australasian Transport Research Forum 2016 Proceedings, Melbourne, Australia, November, 16 – 18, 2016, 2016.
28. [28] H. K. An et al., "Optimized Intersection Signal Timing: An Intelligent Approach-Based Study for Sustainable Models," Sustainability, vol. 14, no. 18, p. 11422, 2022, doi: 10.3390/su141811422.
29. [29] Akçelik & Associates. "SIDRA Glossary of Road Traffic Analysis Terms." https://www.sidrasolutions.com/learn/glossary (accessed Jan 20, 2026).
30. [30] E. Doğan and A. P. Akgüngör, "Optimizing a fuzzy logic traffic signal controller via the differential evolution algorithm under different traffic scenarios," Simulation, vol. 92, no. 11, pp. 1013-1023, 2016, doi: 10.1177/0037549716673217.
31. [31] E. Korkmaz and A. P. Akgüngör, "Optimizing of phase plan, sequence and signal timing based on flower pollination algorithm for signalized intersections," Soft Computing, vol. 25, no. 6, pp. 4243-4259, 2020, doi: 10.1007/s00500-020-05438-x.
32. [32] E. Korkmaz and A. P. Akgüngör, "Comparison of artificial bee colony and flower pollination algorithms in vehicle delay models at signalized intersections," Neural Computing and Applications, vol. 32, no. 8, pp. 3581-3597, 2018, doi: 10.1007/s00521-018-3670-3.
33. [33] Z. Cakici and G. Aksoy, "Does the minimization of the average vehicle delay and the minimization of the average number of stops mean the same at the signalized intersections?," International Journal of Transportation Science and Technology, vol. 13, pp. 213-228, 2024, doi: 10.1016/j.ijtst.2024.01.003.
34. [34] A. A. Assolie, N. S. A. Sukor, I. Khliefat, and T. S. B. Abd Manan, "Modeling of Queue Detector Location at Signalized Roundabouts via VISSIM Micro-Simulation Software in Amman City, Jordan," Sustainability, vol. 15, no. 11, p. 8451, 2023, doi: 10.3390/su15118451.
35. [35] T. R. Currin, Introduction to Traffic Engineering: A Manual for Data Collection and Analysis, 2nd Edition ed. Cengage Learning, 2012.
36. [36] J. Wu, X. Jin, and A. J. Horowitz, "Methodologies for Estimating Vehicle Queue Length at Metered On-Ramps," Transportation Research Record: Journal of the Transportation Research Board, vol. 2047, no. 1, pp. 75-82, 2008, doi: 10.3141/2047-09.
37. [37] A. Bozkurt and P. Çalışkanelli, "Investigation The Effects of One Way/Two Way Applications on Signalized Intersection Performance," presented at the 2. International European Conference on Interdisciplinary Scientific Researches, Ankara, July 4-5, 2020.
38. [38] A. Bozkurt, "İzmir Kent Merkezindeki Trafik Yönetimi Uygulamalarının Başarımlarının İrdelenmesi," Master Thesis, Civil Engineering, Transportation, Dokuz Eylul University, Graduate School of Natural and Applied Sciences, 2021.
39. [39] S. Mosharafian, S. Afzali, and J. Mohammadpour Velni, "Leveraging autonomous vehicles in mixed-autonomy traffic networks with reinforcement learning-controlled intersections," Transportation Letters, pp. 1-12, 2022, doi: 10.1080/19427867.2022.2146302.
40. [40] K. Culik, A. Kalasova, P. Fabian, and J. Ondrus, "Enhancing Intersection Efficiency: A Comparative Analysis of Converting Single-Lane Roundabout to Turbo Roundabout," Sensors, vol. 25, no. 22, p. 7002, 2025, doi: 10.3390/s25227002.
41. [41] B. Şişman, G. Sarisoy, and Y. Delice, "Katlı Kavşak Düzenleme Önerilerinde Simülasyon Programı Kullanım Sürecinin İrdelenmesi," Afyon Kocatepe University Journal of Sciences and Engineering, vol. 21, no. 6, pp. 1434-1452, 2021, doi: 10.35414/akufemubid.931834.
42. [42] A. E. Maxwell, M. Sharma, and K. A. Donaldson, "Explainable Boosting Machines for Slope Failure Spatial Predictive Modeling," Remote Sensing, vol. 13, no. 24, p. 4991, 2021, doi: 10.3390/rs13244991.
43. [43] C. Banyong et al., "Machine Learning-Based Analysis of Travel Mode Preferences: Neural and Boosting Model Comparison Using Stated Preference Data from Thailand’s Emerging High-Speed Rail Network," Big Data and Cognitive Computing, vol. 9, no. 6, p. 155, 2025.
44. [44] A. Khattak, J. Zhang, P.-W. Chan, F. Chen, and H. Almujibah, "Explainable Boosting Machine: A Contemporary Glass-Box Strategy for the Assessment of Wind Shear Severity in the Runway Vicinity Based on the Doppler Light Detection and Ranging Data," Atmosphere, vol. 15, no. 1, pp. 1-19, 2024, doi: 10.3390/atmos15010020.
45. [45] X. Zhang, L. Wu, Z. Li, and H. Liu, "A Robust Method to Measure the Global Feature Importance of Complex Prediction Models," IEEE Access, vol. 9, pp. 7885-7893, 2021, doi: 10.1109/ACCESS.2021.3049412.
46. [46] W. Alajali, W. Zhou, S. Wen, and Y. Wang, "Intersection Traffic Prediction Using Decision Tree Models," Symmetry, vol. 10, no. 9, p. 386, 2018, doi: 10.3390/sym10090386.
47. [47] A. Alsahaf, N. Petkov, V. Shenoy, and G. Azzopardi, "A framework for feature selection through boosting," Expert Systems with Applications, vol. 187, p. 115895, 2022, doi: 10.1016/j.eswa.2021.115895.
48. [48] D. M. Vlachogiannis, S. Moura, and J. Macfarlane, "Intersense: An XGBoost model for traffic regulator identification at intersections through crowdsourced GPS data," Transportation Research Part C: Emerging Technologies, vol. 151, p. 104112, 2023, doi: 10.1016/j.trc.2023.104112.
49. [49] X. Li, P. Xu, and Y.-J. Wu, "Pedestrian crossing volume estimation at signalized intersections using Bayesian additive regression trees," Journal of Intelligent Transportation Systems, vol. 26, no. 5, pp. 557-571, 2021, doi: 10.1080/15472450.2021.1933471.
50. [50] D. Herzalla, W. T. Lunardi, and M. Andreoni, "TII-SSRC-23 Dataset: Typological Exploration of Diverse Traffic Patterns for Intrusion Detection," IEEE Access, vol. 11, pp. 118577-118594, 2023, doi: 10.1109/access.2023.3319213.
51. [51] Z. Chen and W. Fan, "A Freeway Travel Time Prediction Method Based on an XGBoost Model," Sustainability, vol. 13, no. 15, p. 8577, 2021, doi: 10.3390/su13158577.
52. [52] X. Zhang and Q. Zhang, "Short-Term Traffic Flow Prediction Based on LSTM-XGBoost Combination Model," Computer Modeling in Engineering & Sciences, vol. 125, no. 1, pp. 95-109, 2020, doi: 10.32604/cmes.2020.011013.
53. [53] L. Zhu, S. Shu, L. Zou, and M. Fernandez-Veiga, "XGBoost-Based Travel Time Prediction between Bus Stations and Analysis of Influencing Factors," Wireless Communications and Mobile Computing, vol. 2022, no. 3504704, pp. 1-25, 2022, doi: 10.1155/2022/3504704.
54. [54] I. Betkier and M. Oszczypała, "A novel approach to traffic modelling based on road parameters, weather conditions and GPS data using feedforward neural networks," Expert Systems with Applications, vol. 245, p. 123067, 2024, doi: 10.1016/j.eswa.2023.123067.
55. [55] X. Lu, C. Chen, R. Gao, and Z. Xing, "Prediction of High-Speed Traffic Flow around City Based on BO-XGBoost Model," Symmetry, vol. 15, no. 1453, pp. 1-17, 2023, doi: 10.3390/sym15071453.
56. [56] A. Korner et al., "Explainable Boosting Machine approach identifies risk factors for acute renal failure," Intensive Care Med Exp, vol. 12, no. 1, p. 55, Jun 14 2024, doi: 10.1186/s40635-024-00639-2.