Comparison of Smeed Regression Model and Modified Smeed Regression Models for Predicting Road Traffic Fatalities in Türkiye

Abstract

Many factors such as road conditions, driver behavior, vehicle characteristics, environmental conditions and their interactions play a role in the occurrence of traffic accidents. However, it is generally not possible to create a comprehensive model that includes all these factors at the same time, nor is it efficient in terms of practical use. Therefore, it is preferred that the models to be developed are both simple and reliable. In accident forecasting studies conducted in Türkiye, basic variables such as population and number of vehicles, which are directly related to the number of road traffic fatalities, are used. Population growth and economic developments in Türkiye lead to a rapid increase in the number of vehicles in traffic, which in turn leads to an increasing density of vehicle traffic on highways. The rapid increase in the number of vehicles and traffic density causes traffic accidents to reach serious levels. Therefore, knowing the accident rates in a country provides a useful tool for their prevention as well as a comprehensive analysis of their causes. The study aims to estimate road traffic fatalities based on population size and number of active vehicles using Smeed regression method and modified Smeed regression method. In this study on accidents in Türkiye, the number of vehicles, population and fatalities were selected as the main parameters in the modeling process and data for the years 2008-2024 were used. In the models, population and number of motor vehicles are used as independent variables, while death count is considered as the dependent variable. When the results obtained are examined, the MAPE value, which is the average absolute error value of the modified Smeed regression model, is 16.44242%, while that of the Smeed regression model is 16.44449%. On the other hand, the R-squared value for the modified Smeed regression model was calculated as 0.21805021488, while that for the Smeed regression model was 0.11544265593. Thus, it has been observed that the modified Smeed regression model is relatively more accurate than the Smeed regression model in estimating the number of fatalities in road traffic accidents in Türkiye.

Keywords

• Road traffic fatalities
• Number of vehicles
• Regression analysis
• Smeed regression method

References

1. Abdelwahab HT, Abdel-Aty MA. 2001. Development of artificial neural network models to predict driver ınjury severity in traffic accident at signalized ıntersection. Transp Res Rec, 1746: 6-13.
2. Akgüngör AP, Doğan E. 2009. An artificial intelligent approach to traffic accident estimation: Model development and application. Transport, 24(2): 135-142.
3. Akgüngör AP, Doğan E. 2010. Farklı yöntemler kullanılarak geliştirilen trafik kaza tahmin modelleri ve analizi. Int J Eng Res Dev, 2(1): 16-22.
4. Anderson TK. 2009. Kernel density estimation and K means clustering to profile road accident hotspots. Accid Anal Prev, 41(3): 359-364.
5. Andreassen DC. 1985. Linking deaths with vehicles and population. Traffic Eng Control, 26(11): 547-549.
6. Balcı M, Gölcük A, Kahramanli H. 2017. İstatistiksel yaklaşımla trafik kazalarındaki ölüm ve yaralanma durumlarının kusurlu unsurlarla ilişkilerinin incelenmesi. Selçuk-Teknik Derg, 16(3): 210-225.
7. Bener A, Ofosu JB. 1991. Road traffic fatalities in saudi arabia. J Intern Assoc Traffic Safety Sci, 15: 35-38.
8. Cenaj E, Dervishi R. 2024. Road accident fatalities forecasting models using smeed’s regression analysis: A case study. European J Eng Technol Res, 9(6): 20-24.

9. Ceylan H, Haldenbilen S. 2005. Genetik algoritma yaklaşımı ile Avrupa Birliği üyeliği sürecinde Türkiye de beklenen ulaşım talebi ve yönetimi üzerine bir yaklaşım. SDÜ Fen Bil Enst Derg, 9(1): 153-159.
10. Chakrobort S, Roy SK. 2005. Traffic accident characteristics of kolkata. Transport Commun Bullettin Asia Pacific, 74: 75-86.
11. Chang L, Wang H. 2006. Analysis of traffic injury severity: an application of non-parametric classification tree techniques. Accid Anal Prev, 38(5): 1019-1027.
12. Chiou YC. 2006. An artificial network-based expert system for appraisal of two-car crash accidents. Accid Anal Prev, 38(4): 777-785.
13. Chong M, Abraham A, Paprzycki M. 2005. Traffic accident analysis using machine learning paradigms. Informatica, 29(1): 89- 98.
14. Delen D, Sharda R, Besson M. 2006. Identifying significant predictors of ınjury severity in traffic accidents using a series of artificial neural networks. Accid Anal Prev, 38(3): 434-444.
15. Dinesh M, Bawa BS. 1985. An analysis of road traffic fatalities in Delhi, India. Accid Anal Prev, 17(1): 33-45.
16. Ersen M, Büyüklü AH, Taşabat SE. 2022. Data mining as a method for comparison of traffic accidents in şişli district of İstanbul. J Contemp Urban Aff, 6(2): 113- 141.
17. Fouracre P, Jacobs GD. 1977. Further research on road accident rate in developing countries. TRL report 270. Transport Res Lab, Crowthorne, Berkshire, pp: 1-11.
18. Ghee C, Silcock D, Astrop A, Jacobs G. 1997. Socio economic aspects of road accidents in developing countries. TRL Report 247. Transport Res Lab, Crowthorne, Berkshire, pp: 1-33.
19. Gündoğdu Ö, Gökdağ M, Yüksel F. 2005. A traffic noise prediction method based on vehicle composition using genetic algorithms. Appl Acoust, 66(7): 799-809.
20. Haldenbilen S, Ceylan H. 2005. Genetic algorithm approach to estimate transport energy demand in Türkiye. Energy Policy, 33(1): 89-98.
21. Hesse CA, Ofosu JB, Lamptey BL. 2014. A regression model for predicting road traffic fatalities in ghana. Open Sci Repository Math, Online (open-access), e23050497. URL: https://ej-eng.org/index.php/ejeng/article/view/3213/1647 (accessed date: April 15, 2025).
22. Jacobs GD, Bardsley M. 1977. Research on Road Accidents in Developing Countries. Traffic Eng Control, 18(4): 166-170.
23. Jiajia L, Jie H, Ziyang L, Hao Z, Chen Z. 2019. Traffic accident analysis based on C4.5 algorithm in WEKA. MATEC Web Conf, 272(10): 1-8.
24. Junus NWM, İsmail MT, Arsad Z. 2015. Predicting penang road accidents ınfluences: Time series regression versus structural time series. Indian J Sci Technol, 8(30): 1-10.
25. Kıyıldı RK. 2017. Türkiye için yapay sinir ağları yöntemi ile trafik kazası tahmini araştırması. In 5th Int Symp Innovative Technol Eng Sci, September 29-30, Baku, Azerbaijan, p:1642-1651.
26. Koren C, Borsos A. 2012. The advantage of late-comers: Analysis of road fatality rates in the Eu member states. Procedia - Social Behav Sci, 48: 2101-2110.
27. Kuşkapan E, Çodur MY. 2022. Trafik kazalarinin siniflandirilmasinda çok katmanli algilayici regresyon ve en yakin komşuluk algoritmalarinin performans analizi. J Polytechnic, 25(1): 373-380.
28. Kwon OH, Rhee W, Yoon Y. 2015. Application of classification algorithms for analysis of road safety risk factor dependencies. Accid Anal Prev, 75: 1-15.
29. Livneh M, Hakkert AS. 1972. Some factors effecting the ıncrease road accidents in developing countries, with particular reference to Israel. Accid Anal Prev, 4: 117-133.
30. Malka L, Bidaj F. 2015. Opacity evaluation for passenger diesel vehicle cars in Tirana. J Environ Sci Eng, 4(7): 352-358.
31. Malka L, Dervishi R, Malkaj P, Konomi I, Ormeni R, Cenaj E. 2024. Modelling and assessing environmental ımpact in transport to meet the sector’s climate goals in 2050. WSEAS Trans Environ Dev, 20: 350-364.
32. Malka L, Konomi I, Bartocci P, Rrapaj E. 2021. An integrated approach toward a sustainable transport sector using EnergyPLAN model: Case of Albania. Innovations, 9(4): 141-147.
33. Mekky A. 1985. Effect of rapid ıncrease in motorization levels on road fatality rates in some rich developing countries. Accid Anal Prev, 17(2): 101-109.
34. Muhammad LJ, Salisu S, Yakubu A, Malgwi YM, Abdullahi ET, Mohammed IA, Muhammad NA. 2017. Using decision tree data mining algorithm to predict causes of road traffic accidents, its prone locations and time along kano-wudil highway. Int J Database Theory Appl, 10(1): 197-206.
35. Mussone L, Ferrari A, Oneta M. 1999. An analysis of urban collision using an artificial ıntelligence model. Accid Anal Prev, 31(8): 705-718.
36. Oña J, Mujalli RO, Calvo FJ. 2011. Analysis of traffic accident injury severity on Spanish rural highways using bayesian networks. Accid Anal Prev, 43(1): 402-411.
37. Özgan E, Ulusu H, Yıldız K. 2004. Trafik kaza verilerinin analizi ve kaza tahmin modeli. SAU Fen Bil Enst Derg, 8(1): 160-166.
38. Partyka C. 1984. Simple models of fatality trends using employment and population data. Accid Anal Prev, 16(3): 211-222.
39. Smeed RJ, Jaffocate GO. 1970. Effects of changes in motorization in various countries on the number of road fatalities. Traffic Eng Control, 12(3): 150-151.
40. Smeed RJ. 1949. Some statistics aspects of road safety research. J R Stat Soc, Series A, Part I: 1-34.
41. Smeed RJ. 1968, Variations in the pattern of accident rates in different countries and their causes. Traffic Eng Control, 10(7): 364-371.
42. Sohn SY, Shin H. 2010. Pattern recognition for road traffic accident severity in Korea. Ergon, 44(1): 107-117.
43. TURKSTAT. 2024a. Highway traffic accident statistics (2008-2024). URL: https://www.tuik.gov.tr (accessed date: March 3, 2025).
44. TURKSTAT. 2024b. Address based population registration system results (2008-2024). URL: https://www.tuik.gov.tr (accessed date: March 4, 2025).
45. Valli PP. 2005. Road accident models for large metropolitan cities of India. IATSS Research, 29 (1): 57-65.
46. WHO. 2023. Global status report on road safety (2023). URL: https://iris.who.int/bitstream/handle/10665/375016/9789240086517-eng.pdf?sequence=1 (accessed date: August 5, 2024).
47. Yavuz AA, Ergül B, Aşik EG. 2021. Trafik kazalarının makine öğrenmesi yöntemleri kullanılarak değerlendirilmesi. Int J Eng Res Dev, 13(1): 66-73.
48. Yılmaz MB, Çilengiroğlu ÖV. 2022. Talep tahminleme değişkenlerinin üssel düzeltme yöntemi ile belirlenmesi. Euroasia J Math Eng Nat Med Sci, 9(22): 92-103.
49. Zegeer CV, Deacon JA. 1987. Effect of lane width, shoulder width, and shoulder type on highway safety. In: Relationship between safety and key highway features. State of the Art Report 6, Transp Res Board, Washington, US, pp: 1-21.
50. Zengı̇n B, Kaymaz K, Arslannur B. 2018. Tunceli ilindeki trafik kazası oranlarının incelenmesi. Gümüşhane Üniv Fen Bil Enst Derg, 8(2): 318-324.