İnşaat sektöründe standart ve düzenlileştirilmiş lojistik regresyon modelleri kullanılarak iş kazalarının şiddetinin tahmini

Year 2023, , 778 - 798, 15.07.2023

Şura Toptancı , Nihal Erginel , Ilgın Poyraz Acar

https://doi.org/10.28948/ngumuh.1212385

Abstract

İnşaat sektöründe iş kazaları diğer sektörlere kıyasla daha sık meydana gelmektedir. İnşaat iş kazaları hâlâ istenilen düzeyde önlenememiştir. Literatürde klasik istatistiksel ve makine öğrenmesi teknikleri kullanılarak bu kazaların meydana gelme sıklığını tahmin etmek için birçok çalışma yapılmaktadır. Ancak, büyük boyutlu ve çok sayıda kategorik değişken içeren veriler analiz edilirken, veri setinde bulunan dengesizlik ve çoklu bağlantı sorunlarına ilişkin bazı problemler dikkate alınmamaktadır. Bu çalışma daha doğru sonuçlar elde edebilmek için bahsedilen problemleri dikkate alarak, ölümcül olmayan inşaat kazalarının şiddetini tahmin etmeyi amaçlamaktadır. Bu çalışmada, inşaat sektöründe iş günü kaybının tahmini için standart ikili lojistik regresyon, Firth, Ridge, Lasso ve Elastik Net düzenlileştirilmiş lojistik regresyon modelleri kullanılmış ve sonuçlar karşılaştırılmıştır. Kullanılan veriler kazazede, iş yeri, kaza zamanı, kaza ve olaylar zinciri ve kaza sonrası durumla ilgili değişkenler olmak üzere beş gruba ayrılmıştır. Sonuçlar, Firth'in lojistik modelinin en iyi performans gösteren model olduğunu ve yaş, eğitim, mesleki eğitim, işyeri büyüklüğü, proje türü, çalışılan ortam, kaza ayı ve yılı, genel ve özel faaliyetler, kullanılan materyal, yaranın türü ve yaranın vücuttaki yerinin en önemli değişkenler olduğunu göstermiştir. Yorumlanabilir makine öğrenimi araçları sağlayan bu çalışma, literatürde inşaat güvenliği alanında önerilen modelleri kullanmaya yönelik ilk girişimdir.

Keywords

İş kazaları, İnşaat sektörü, Lojistik regresyon, Makine öğrenmesi, Kaza şiddeti

Supporting Institution

Eskişehir Teknik Üniversitesi Bilimsel Araştırma Projeleri Komisyonu (1705F427 nolu proje). Bu çalışma, birinci yazarın hazırladığı doktora tezinden üretilmiştir.

Project Number

BAP 1705F427

Thanks

Yazarlar, çalışmada kullanılan verinin temini için katkı sağlayan Sosyal Güvenlik Kurumu ve Hizmet Sunumu Genel Müdürlüğü’ne teşekkür eder.

Predicting the severity of occupational accidents in the construction industry using standard and regularized logistic regression models

Abstract

Occupational accidents in the construction industry occur more frequently when compared with other industries. Construction occupational accidents still have not been prevented at the desired level. Several studies in the literature have been conducted to predict the occurrence frequency of these accidents using classical statistical and machine-learning techniques. However, some challenges regarding imbalanced and multicollinearity problems present in the dataset are not considered while analyzing data with a large size and a large number of categorical variables. This study aims to predict the severity of non-fatal construction accidents considering mentioned challenges to obtain more accurate results. In this study, standard binary logistic regression, Firth, Ridge, Lasso, and Elastic Net Regularized logistic regression models were used for the prediction of lost workdays in the construction industry and results were compared. The data used were classified into five groups: victim, workplace, accident time, accident and sequence of events, and post-accident state-related variables. The results showed that Firth’s logistic model is the best-performing model and age, education, vocational education, workplace size, project type, working environment, accident month and year, general and specific activities, material agent, type of injury, and part of body injured are the most significant variables. This study, by providing interpretable machine learning tools, is the first attempt to use proposed models in the area of construction safety in the literature.

Keywords

Occupational accidents, Construction industry, Logistic regression, Machine learning, Accident severity

Project Number

BAP 1705F427

References

• J.P. Leigh, S.B. Markowitz, M. Fahs and P. Landrigan, Costs of Occupational Injuries and Illnesses. University of Michigan Press, Ann Arbor, Mich., 2000.
• L.I. Boden, E.A. Biddle and E.A. Spieler, Social and economic impacts of workplace illness and injury: current and future directions for research. American Journal of Industrial Medicine, 40, 398-402, 2001. https://doi.org/10.1002/ajim.10013.
• S. Linacre, Australian social trends 2007. Australian Bureau of Statistics. ABS catalogue no. 4102. https://www.abs.gov.au/AUSSTATS/abs@.nsf/allprimarymainfeatures/3550D34DA999401ECA25748E00126282?opendocument, Accessed 01 March 2021.
• D.L. Goetsch, Occupational Safety and Health for Technologists, Engineers, and Managers. 6th ed. Pearson Prentice Hall, 2008.
• International Labour Organization (ILO), Databases and subjects: labour force by sex and age-2020. https://ilostat.ilo.org/topics/population-and-labour-force/#, Accessed 23 May 2021.
• International Labour Organization (ILO), Safety and health at work. https://www.ilo.org/global/ topics/safety-and-health-at-work/lang--en/index.htm , Accessed 5 December 2020.
• P. Hämäläinen, J. Takala and T.B. Kiat, Global estimates of occupational accidents and work-related illnesses 2017. https://www.icohweb.org/site/images/ news/pdf/Report%20Global%20Estimates%20of%20Occupational%20Accidents%20and%20Work-related%20Illnesses%202017%20rev1.pdf, Accessed 5 December 2020.
• Bureau of Labor Statistics (BLS), Census of fatal occupational injuries - 2018. https://www.bls.gov/ iif/oshcfoi1.htm, Accessed 7 December 2020.
• SGK, SGK İstatistik Yıllıkları- 2018. http://www.sgk.gov.tr/wps/portal/sgk/tr/kurumsal/istatistik/sgk_istatistik_yilliklari, 8 December 2020.
• H. Cakan, Analysis and modeling of roofer and steel worker fall accidents. Ph.D. Thesis, Wayne State University, Michigan, USA, 2012.
• S. Onder, Evaluation of occupational injuries with lost days among opencast coal mine workers through logistic regression models. Safety Science, 59, 86-92, 2013, http://dx.doi.org/10.1016/j.ssci.2013.05.002.
• Ö. Akboga, Modeling of construction accident severity using logistic regression. Ph.D. Thesis, Ege University, İzmir, Türkiye, 2014.
• A. Bilim, Analysis and modeling of occupational accidents occurring in highway and railway constructions. Ph.D. Thesis, Konya Technic University, Konya, Türkiye, 2018.
• A.J. Tixier, M.R. Hallowell, B. Rajagopalan and D. Bowman, Application of machine learning to construction injury prediction. Automation in Construction, 69, 102-114, 2016. https://doi.org/10.1016/j.autcon.2016.05.016.
• K. Yang, C.R. Ahn, M.C. Vuran and S.S. Aria, Semi-supervised near-miss fall detection for ironworkers with a wearable inertial measurement unit. Automation in Construction, 68, 194-202, 2016. https://doi.org/10.1016/j.autcon.2016.04.007.
• K. Kang and H. Ryu, Predicting types of occupational accidents at construction sites in Korea using random forest model. Safety Science, 120:226-236, 2019. https://doi.org/10.1016/j.ssci.2019.06.034.
• B.U. Ayhan and O.B. Tokdemir, Safety assessment in megaprojects using artificial intelligence. Safety Science, 118:273-287, 2019. https://doi.org/ 10.1016/j.ssci.2019.05.027.
• J.Y. Lee, Y.G. Yoon, T.K. Oh, S. Park and S.I. Ryu, A study on data pre-processing and accident prediction modelling for occupational accident analysis in the construction industry. Applied Sciences, 10(21), 7949, 2020. https://doi.org/10.3390/app10217949.
• J. Choi, B. Gu, S. Chin and J.S. Lee, Machine learning predictive model based on national data for fatal accidents of construction workers. Automation in Construction, 110, 102974, 2020. https://doi.org/ 10.1016/j.autcon.2019.102974.
• F. Recal and T. Demirel, Comparison of machine learning methods in predicting binary and multi-class occupational accident severity. Journal of Intelligent & Fuzzy Systems, 40(6), 10981-10998, 2021. https://doi.org/10.3233/JIFS-202099.
• Y.Ö. Tetik, Ö. Akboğa Kale, I. Bayram and S. Baradan, Applying decision tree algorithm to explore occupational injuries in the Turkish construction industry. Journal of Engineering Research, 10(3), 59-70, 2022. https://doi.org/10.36909/jer.12209.
• K. Koc, Ö. Ekmekcioğlu and A.P. Gurgun, Prediction of construction accident outcomes based on an imbalanced dataset through integrated resampling techniques and machine learning methods. Engineering, Construction and Architectural Management, (ahead-of-print), 2022. https://doi.org/ 10.1108/ECAM-04-2022-0305.
• J.M. Pereira, M. Basto and A.F. da Silva, The logistic lasso and ridge regression in predicting corporate failure. Procedia Economics and Finance, 39, 634-641, 2016. https://doi.org/10.1016/S2212-5671(16)30310-0.
• R. Gavanji, Penalized regression methods for modelling rare events data with application to occupational injury study. Master Thesis, University of Saskatchewan, Canada, 2019.
• S. Sarkar and J. Maiti, Machine learning in occupational accident analysis: A review using science mapping approach with citation network analysis. Safety science, 131, 104900, 2020. https://doi.org/10.1016/j.ssci.2020.104900.
• M. Gonzalez-Delgado, H. Gómez-Dantés, J.A. Fernández-Niño, E. Robles, V.H. Borja and M. Aguilar, Factors associated with fatal occupational accidents among Mexican workers: a national analysis. PloS one, 2015. https://doi.org/10.1371/journal.pone.0121490.
• S.S. Uysal, Comparison of The Logistic Elastic Net Method with Alternative Methods. Master Thesis, Eskisehir Osmangazi University, Türkiye, 2020.
• V. Gallego, A. Sánchez, I. Martón and S. Martorell, Analysis of occupational accidents in Spain using shrinkage regression methods. Safety Science, 133, 105000, 2021. https://doi.org/10.1016/j.ssci.2020.105000.
• D.W. Hosmer, S. Lemeshow and R.X. Sturdivant, Applied Logistic regression. 3rd ed. Hoboken, New Jersey: Wiley, 2013.
• A. Agresti, An introduction to categorical data analysis. 3rd ed. Hoboken, NJ: John Wiley & Sons, 2019.
• S.A. Czepiel, Maximum likelihood estimation of logistic regression models: theory and implementation. czep. net/stat/mlelr.pdf. Accessed 20 August 2022.
• B.G. Tabachnick and L.S. Fidell, Using multivariate statistics. 6th ed. Boston: Pearson, 2013.
• G. Kemalbay and B.N. Alkış, Borsa endeks hareket yönünün çoklu lojistik regresyon ve k-en yakın komşu algoritması ile tahmini. Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi, 27(4), 556-569, 2020. https://doi.org/10.5505/pajes.2020.57383.
• D.N. Gujarati and D.C. Porter, Basic econometrics. 5th ed. New York: McGraw-Hill/Irwin, 2009.
• G. King and L. Zeng, Logistic regression in rare events data. Political Analysis, 9, 137-163, 2001. https://doi.org/10.1093/oxfordjournals.pan.a004868.
• C.F. İşçen, S.S. Uysal and A.A. Yavuz, Su kalitesi değişimine etki eden değişkenlerin lojistik regresyon, lojistik-ridge ve lojistik lasso yöntemleri ile tespiti. Biyoloji Bilimleri Araştırma Dergisi, 14(1), 1-12, 2021. https://bibad.gen.tr/ index.php/bibad/article/ view/375.
• Z.Y. Algamal and M.H. Lee, Applying penalized binary logistic regression with correlation based elastic net for variables selection. Journal of Modern Applied Statistical Methods, 14(1), 168-179, 2015. https://doi.org/10.22237/jmasm/1430453640.
• S. Doerken, M. Avalos, E. Lagarde and M. Schumacher, Penalized logistic regression with low prevalence exposures beyond high dimensional settings. PLoS One, 14(5), e0217057, 2019. https://doi.org/10.1371/journal.pone.0217057.
• D. Firth, Bias reduction of maximum likelihood estimates. Biometrika, 80, 27–38, 1993. https://doi.org/10.2307/2336755.
• M.S. Rahman and M. Sultana, Performance of Firth-and logF-type penalized methods in risk prediction for small or sparse binary data. BMC Medical Research Methodology, 17(1), 1-15, 2017. https://doi.org/ 10.1186/s12874-017-0313-9.
• R.L. Schaefer, L.D. Roi and R.A. Wolfe, A ridge logistic estimator. Communications in Statistics -Theory and Methods, 13(1), 99-113, 1984. https://doi.org/10.1080/03610928408828664.
• D.E. Duffy and T.J. Santne, On the Small Sample Properties of Norm-Restricted Maximum Likelihood Estimators for Logistic Regression Models. Communications in Statistics - Theory and Methods, 18, 959-980, 1989. https://doi.org/ 10.1080/03610928908829944.
• S. Le Cessie and J.C. Van Houwelingen, Ridge estimators in logistic regression. Journal of the Royal Statistical Society: Series C (Applied Statistics), 41(1), 191-201, 1992. https://doi.org/10.2307/2347628.
• W.J. Fu, Penalized regressions: the bridge versus the lasso. Journal of Computational and Graphical Statistics, 7(3), 397-416, 1998. https://doi.org/ 10.2307/1390712.
• G. James, D. Witten, T. Hastie, and R. Tibshirani, Linear model selection and regularization. In: An Introduction to Statistical Learning, Springer Text in Statistics, , 225-288, Springer, New York, NY, 2021.
• R. Tibshirani, Regression Shrinkage and Selection via the LASSO. Journal of Royal Statistical Society B, 58, 267–288, 1996. https://doi.org/10.1111/j.2517-6161.1996.tb02080.x.
• H. Zou and T. Hastie, Regularization and variable selection via the elastic net. Journal of the royal statistical society: series B (statistical methodology), 67(2), 301– 320, 2005. https://doi.org /10.1111/j.1467-9868.2005.00503.x.
• T. Hastie, R. Tibshirani and M. Wainwright, Statistical learning with sparsity. Monographs on statistics and applied probability, 143, CRC Press, 2015.
• B. Jason, A Gentle Introduction to k-fold Cross-Validation. https: //machinelearningmastery.com/k-fold-cross-validation/, Accessed 1 August 2021.
• T. Hastie, R. Tibshirani, and J.H. Friedman, The elements of statistical learning: data mining, inference, and prediction. New York: Springer, 2001.
• M. Grandini, E. Bagli and G. Visani, Metrics for multi-class classification: an Overview. A white paper, https://arxiv.org/pdf/2008.05756.pdf, Accessed 8 May 2021.
• T. Saito and M. Rehmsmeier, The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE, 10(3), e0118432, 2015. https://doi.org/10.1371/journal.pone.0118432.
• D. Chicco, N. Tötsch and G. Jurma, The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix. BioData Mining, 14(13), 1-22, 2021. https://doi.org/ 10.1186/s13040-021-00244-z.
• H. Guo, H. Liu, C. Wu, W. Zhi, Y. Xiao and W. She, Logistic discrimination based on G-mean and F-measure for imbalanced problem. Journal of Intelligent & Fuzzy Systems, 31(3), 1155-1166, 2016. https://doi.org/10.3233/IFS-162150.
• J. Davis and M. Goadrich, The relationship between Precision-Recall and ROC curves. In Proceedings of the 23rd international conference on Machine learning, pp. 233-240, Pittsburgh, PA, 2006.
• H.R. Sofaer, J.A. Hoeting and C.S. Jarnevich, The area under the precision‐recall curve as a performance metric for rare binary events. Methods in Ecology and Evolution, 10(4), 565-577, 2019. https://doi.org/ 10.1111/2041-210X.13140.
• Eurostat. European Statistics on Accidents at Work (ESAW), Summary methodology 2013. https://ec.europa.eu/eurostat/documents/3859598/5926181/KS-RA-12-102-EN.PDF/56cd35ba-1e8a-4af3-9f9a-b3c47611ff1c, Accessed 1 August 2020.
• W. Jiang, P. Lakshminarayanan, X. Hui, P. Han, Z. Cheng, M. Bowers, I. Shpitser, S. Siddiqui, R.H. Taylor, H. Quon and T. McNutt, Machine learning methods uncover radiomorphologic dose patterns in salivary glands that predict xerostomia in patients with head and neck cancer. Advances in radiation oncology, 4(2), 401-412, 2019. https://doi.org/ 10.1016/j.adro.2018.11.008.
• Ş. Toptancı, Prediction of occupational accidents, risk analysis and prioritization of precautions to be taken by house of quality in construction industry (Doctoral dissertation). Available from Turkish Council of Higher Education Thesis Center (Thesis No: 694456), 2021.