A comparative analysis of machine learning algorithms for estimating Annual Average Daily Traffic (AADT) on provincial roads in Turkey

Year 2025, Volume: 14 Issue: 4

Serkan Biçici

Abstract

Annual Average Daily Traffic (AADT) is the average number of vehicles passing on a road over the course of a year. In Turkey, AADT data is collected by the General Directorate of Highways for three different road classes. For state roads and highways, AADT data is collected regularly every year, but for provincial roads, this data is usually collected every three or four years. In this study, seven different machine learning algorithms were used to estimate AADT for provincial roads and their performances were compared. Variables reflecting the characteristics of the road, its relationship with other transportation systems and the demographic/socio-economic characteristics of the road environment were used. Random forest and support vector regression algorithms produced the most successful results. Variables representing road characteristics and demographic/socio-economic factors were found to have a much more significant impact on AADT estimates than variables reflecting the relationship with other transportation systems.

Keywords

AADT , Provincial Roads , Machine Learning , Random Forest , Support Vecot Regression

Project Number

Artvin Çoruh Üniversitesi Bilimsel Araştırma Projeleri Koordinatörlüğü (Proje No: 2024.F40.02.02).

Türkiye’deki il yolları için Yıllık Ortalama Günlük Trafik (YOGT) tahmininde makine öğrenmesi algoritmalarının karşılaştırmalı analizi

Abstract

Yıllık Ortalama Günlük Trafik (YOGT), bir yoldan bir yıl boyunca geçen ortalama araç sayısını ifade etmektedir. Türkiye’de YOGT verileri Karayolları Genel Müdürlüğü tarafından üç farklı idari yol sınıfı için toplanmaktadır. Devlet ve otoyollar için YOGT verisi her yıl düzenli olarak elde edilmekte, ancak il yolları için bu veri genellikle üç veya dört yılda bir toplanmaktadır. Bu çalışmada, il yolları için YOGT tahmini amacıyla yedi farklı makine öğrenmesi algoritması kullanılmış ve performansları karşılaştırılmıştır. Yolun karakteristik özellikleri, diğer ulaşım sistemleriyle ilişkisini ve yol çevresinin demografik/sosyo-ekonomik özelliklerini yansıtan değişkenler kullanılmıştır. Rastgele orman ve destek vektör regresyonu algoritmaları en başarılı sonuçları vermiştir. Yolun karakteristik özellikleri ile demografik/sosyo-ekonomik faktörleri temsil eden değişkenlerin YOGT tahminleri üzerindeki etkisinin, diğer ulaşım sistemleri merkezleriyle olan ilişkileri yansıtan değişkenlere kıyasla çok daha belirleyici olduğu bulunmuştur.

Keywords

YOGT , İl Yolları , Makine Öğrenimi , Rastgele Orman , Destek Vektör Regresyonu

Project Number

Artvin Çoruh Üniversitesi Bilimsel Araştırma Projeleri Koordinatörlüğü (Proje No: 2024.F40.02.02).

Thanks

Bu çalışma, Artvin Çoruh Üniversitesi Bilimsel Araştırma Projeleri Koordinatörlüğü (AÇÜBAP) tarafından desteklenmiştir (Proje No: 2024.F40.02.02). Ayrıca, bu çalışma büyük ölçüde, hem internet üzerinden hem de özel talepler yoluyla erişilen devlet belgelerine dayanmaktadır. Bu belgelerin temin edilmesindeki değerli yardımları için KGM ve TUCBS personeline en içten teşekkürlerimi sunarım.

References

• R. P. Roess, E. S. Prassas and W. R. McShane, Traffic engineering. Pearson/Prentice Hall, 2004.
•   A. Sfyridis and P. Agnolucci, Annual average daily traffic estimation in England and Wales: An application of clustering and regression modelling. Journal of Transport Geography, 83, 102658, 2020. https://doi.org/10.1016/j.jtrangeo.2020.102658.
•   X. Zhang and M. Chen, Enhancing statewide annual average daily traffic estimation with ubiquitous probe vehicle data. Transportation Research Record, 2674(9), 649-660, 2020. https://doi.org/10.1177/0361198120931100.
•   Ulaştırma ve Altyapı Bakanlığı - Karayolu Genel Müdürlüğü, Trafik ve Ulaşım Bilgileri. https://www.kgm.gov.tr/Sayfalar/KGM/SiteTr/Trafik/TrafikveUlasimBilgileri.aspx, Accessed 01 May 2025.
•   D. Mohamad, K. C. Sinha, T. Kuczek and C. F. Scholer, Annual average daily traffic prediction model for county roads. Transportation research record, 1617(1), 69-77, 1998. https://doi.org/10.3141/1617-10.
•   Q. Xia, F. Zhao, Z. Chen, L. D. Shen and D. Ospina, Estimation of annual average daily traffic for nonstate roads in a Florida county. Transportation Research Record, 1660(1), 32-40, 1999. https://doi.org/10.3141/1660-05.
•   L. Pun, P. Zhao and X. Liu, A multiple regression approach for traffic flow estimation. IEEE access, 7, 35998-36009, 2019. https://doi.org/10.1109/ACCESS.2019.2904645.
•   F. Zhao and N. Park, Using geographically weighted regression models to estimate annual average daily traffic. Transportation research record, 1879(1), 99-107, 2004. https://doi.org/10.3141/1879-12.
•   B. Selby and K. M. Kockelman, Spatial prediction of traffic levels in unmeasured locations: applications of universal kriging and geographically weighted regression. Journal of Transport Geography, 29, 24-32, 2013. https://doi.org/10.1016/j.jtrangeo.2012.12.009.
• M. Doustmohammadi and M. Anderson, A Bayesian regression model for estimating average daily traffic volumes for low volume roadways. International Journal of Statistics and Probability, 8(1), 143-149, 2019. https://doi.org/10.5539/ijsp.v8n1p143.
• D. Apronti, K. Ksaibati, K. Gerow and J. J. Hepner, Estimating traffic volume on Wyoming low volume roads using linear and logistic regression methods. Journal of traffic and transportation engineering (English edition), 3(6), 493-506, 2016. https://doi.org/10.1016/j.jtte.2016.02.004.
• K. Wang, R. DeVine, N. Huynh, W. Jin, G. Comert and M. Chowdhury, Comparison of models with and without roadway features to estimate annual average daily traffic at non-coverage locations. International Journal of Transportation Science and Technology, 15, 244-259, 2024. https://doi.org/10.1016/j.ijtst.2023.10.001.
• X. Sun and S. Das, Developing a method for estimating AADT on all Louisiana roads. Louisiana Transportation Research Center, Louisiana, USA, Technical Report FHWA/LA.14/548, July 2015.
• F. Harrou, A. Zeroual and Y. Sun, Traffic congestion monitoring using an improved kNN strategy. Measurement, 156, 107534, 2020. https://doi.org/10.1016/j.measurement.2020.107534.
• D. Xu, Y. Wang, P. Peng, S. Beilun, Z. Deng and H. Guo, Real-time road traffic state prediction based on kernel-KNN. Transportmetrica A: Transport Science, 16(1), 104-118, 2020. https://doi.org/10.1080/23249935.2018.1491073.
• Y. Zhang and A. Haghani, A gradient boosting method to improve travel time prediction. Transportation Research Part C: Emerging Technologies, 58, 308-324, 2015. https://doi.org/10.1016/j.trc.2015.02.019.
• Y. Han, T. Peng, C. Wang, Z. Zhang and G. Chen, A hybrid GLM model for predicting citywide spatio-temporal metro passenger flow. ISPRS International Journal of Geo-Information, 10(4), 222, 2021. https://doi.org/10.3390/ijgi10040222.
• B. K. Koo, J. W. Baek and K. Y. Chung, Weight feedback-based harmonic MDG-ensemble model for prediction of traffic accident severity. Applied Sciences, 11(11), 5072, 2021. https://doi.org/10.3390/app11115072.
• Türkiye Ulusal Coğrafi Bilgi Sistemleri, Ulasal Coğrafi Bilgi Platformu. https://tucbs.gov.tr, Accessed 01 May 2025.
• Türkiye İstatistik Kurumu (TUİK). https://tuik.gov.tr, Accessed 01 May 2025.
• J. Li, J. Boonaert, A. Doniec and G. Locenguez, Multi-models machine learning methods for traffic flow estimation from floating car data. Transportation Research Part C: Emerging Technologies, 132, 103389, 2021. https://doi.org/10.1016/j.trc.2021.103389.
• İ. Bütüner, B. Kaplan ve K. Adem, Rseslıbknn makine öğrenmesi yöntemi kullanılarak Parkinson Hastalığının Tanısı. Niğde Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi, 9(2), 715-721, 2020. https://doi.org/10.28948/ngumuh.655720.
• S. Zeybek, Derin öğrenme ve makine öğrenmesi yöntemleri ile sosyal medya verilerinden suç tespiti. Niğde Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi, 14(1), 175-182, 2025. https://doi.org/10.28948/ngumuh.1551734.
• Y. Dokuz, A. Bozdağ ve B. Gökçek, Hava kalitesi parametrelerinin tahmini ve mekansal dağılımı için makine öğrenmesi yöntemlerinin kullanılması. Niğde Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi, 9(1), 37-47, 2020. https://doi.org/10.28948/ngumuh.654092.
• Y. M. Kızılkaya ve A. Oğuzlar, Bazı denetimli öğrenme algoritmalarının R programlama dili ile kıyaslanması. Karadeniz Uluslararası Bilimsel Dergi, 37, 90-98, 2018. https://doi.org/10.17498/kdeniz.405746.
• T. Jiang, J. L. Gradus and A. J. Roselline, Supervised machine learning: a brief primer. Behavior therapy, 51(5), 675-687, 2020. https://doi.org/10.1016/j.beth.2020.05.002.
• T. K. Ho, Random decision forests. In: Proceedings of 3rd international conference on document analysis and recognition IEEE, pp. 278-282, Montreal, Canada, 1995
• L. Breiman, Random forests. Machine learning, 45, 5-32,2001. https://doi.org/10.1023/A:1010933404324.
• D. Basak, S. Pal and D. C. Patranabis, Support vector regression. Neural Information Processing-Letters and Reviews, 11(10), 203-224, 2007.
• X. Luo, D. Li, Y. Yang and S. Zhang, Spatiotemporal traffic flow prediction with KNN and LSTM. Journal of Advanced Transportation, 2019(1), 4145353, 2019. https://doi.org/10.1155/2019/4145353.
• I. L. Cherif and A. Kortebi, On using extreme gradient boosting (XGBoost) machine learning algorithm for home network traffic classification. In: 2019 Wireless Days (WD), pp. 1-6, Manchester, UK, 2019. https://doi.org/10.1109/WD.2019.8734193.
• A. J. Dobson and A. G. Barnett, Generalized linear models. Routledge, New York, 2019.
• B. Mahesh, Machine learning algorithms-a review. International Journal of Science and Research, 9(1), 381-386, 2020. https://doi.org/10.21275/ART20203995.
• A. Singh, N. Thakur and A. Sharma, A review of supervised machine learning algorithms. In: 2016 3rd international conference on computing for sustainable global development (INDIACom), pp. 1310-1315, New Delhi, India, 2016.
• G. Bonaccorso, Machine Learning Algorithms: Popular algorithms for data science and machine learning. Packt Publishing Ltd, 2018.
• G. James, D. Witten, T. Hastie, and R. Tibshirani, An introduction to statistical learning: with applications in R. New York: springer, 2013.
• J. H. Friedman, Greedy function approximation: a gradient boosting machine. Annals of statistics,29(5), 1189-1232, 2001.
• K. Max, W. Jed, W. Steve, W. Andre, K. Chris, E. Allan, C. Tony, M. Zachary and et.al. Package ‘caret’. The R Journal, 223.7: 48, 2020.
• P. Probst, A. L. Boulesteix, and B. Bischl, Tunability: Importance of hyperparameters of machine learning algorithms. Journal of Machine Learning Research, 20(53), 1-32, 2019.
• A. Bagnall and G. C. Cawley, On the use of default parameter settings in the empirical evaluation of classification algorithms. arXiv preprint arXiv:1703.06777, 2017.
• R. G. Mantovani, A. L. Rossi, J. Vanschoren, B. Bischl and A. C. Carvalho, To tune or not to tune: Recommending when to adjust SVM hyper-parameters via meta-learning. In 2015 International joint conference on neural networks (IJCNN), pp. 1-8, Killarney, Ireland, 2015.
• S.S. Pulugurtha and P.R. Kusam, Modeling annual average daily traffic with integrated spatial data from multiple network buffer bandwidths. Transportation research record, 2291(1), 53-60, 2012. https://doi.org/10.3141/2291-07.