Random Forest–Based Covariate Shift in Addressing Nonstationarity of Railway Track Data
Publication: ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering
Volume 7, Issue 3
Abstract
The multimodal transportation network in which freight rail plays an essential role continues to enhance the United States’ contributions in the global market. For years, track geometry defects data have been often gathered by visual inspections. However, automated track vehicles are now deployed for the same purpose to save time and cost. One of the limitations of an automated vehicle is the likelihood of nonstationarity of the gathered data due to external influence. The effect of nonstationarity may lead to the wrong representation of track conditions and thereby increases the possibility of false model output. This study applies supervised machine learning (ML) methods to detect the nonstationarity of the geometry data. The methods include random forest (RF), logistic regression (LR), and support vector machine (SVM). The authors vary the train test and validation ratio in phases to ascertain each machine learning method’s accuracy and adaptability to different distributions. In the first phase, the random forest and the support vector machine show an accuracy of 97.1%, while the logistic regression reveals 96% accuracy. In the second and third phases, the random forest method gives a better result than other supervised learners with accuracies of 97% and 97.1%, respectively. Similarly, for validation, the random forest performs better than other classifiers. Conclusively, the developed models’ application indicates that the random forest is a more effective approach to detecting the nonstationarity of track geometry data.
Get full access to this article
View all available purchase options and get full access to this article.
Data Availability Statement
The original data used for the analysis are available in CSV files. The models developed using Jupyter Notebook are available and can be used to check the output of the analysis. The Jupyter Notebook contains both the formulation and code used in the analysis. All these are available from the corresponding author upon reasonable request.
References
Andrade, A. R., and P. F. Teixeira. 2015. “Statistical modelling of railway track geometry degradation using hierarchical Bayesian models.” Reliab. Eng. Syst. Saf. 142 (Oct): 169–183. https://doi.org/10.1016/j.ress.2015.05.009.
Association of American Railroads. 2020. Railroad 101. Washington, DC: Association of American Railroads.
Bai, L., R. Liu, Q. Sun, F. Wang, and P. Xu. 2015. “Markov-based model for the prediction of railway track irregularities.” Proc. Inst. Mech. Eng. 229 (2): 150–159. https://doi.org/10.1177/0954409713503460.
Berawi, A. R. B., R. Delgado, R. Calçada, and C. Vale. 2010. “Evaluating track geometrical quality through different methodologies.” Int. J. Technol. 1 (1): 38–47. https://doi.org/10.14716/ijtech.v1i1.1000.
Cárdenas-Gallo, I., C. A. Sarmiento, G. A. Morales, M. A. Bolivar, and R. Akhavan-Tabatabaei. 2017. “An ensemble classifier to predict track geometry degradation.” Reliab. Eng. Syst. Saf. 161 (Apr): 53–60. https://doi.org/10.1016/j.ress.2016.12.012.
Chenariyan Nakhaee, M., D. Hiemstra, M. Stoelinga, and M. van Noort. 2019. “The recent applications of machine learning in rail track maintenance: A survey.” In Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics), 91–105. Berlin: Springer.
Faghih-Roohi, S., S. Hajizadeh, A. Núñez, R. Babuska, and B. De Schutter. 2016. “Deep convolutional neural networks for detection of rail surface defects.” In Proc., Int. Joint Conf. on Neural Networks, 2584–2589. New York: IEEE.
Falamarzi, A., S. Moridpour, and M. Nazem. 2019. “A review of rail track degradation prediction models.” Aust. J. Civ. Eng. 17 (2): 152–166. https://doi.org/10.1080/14488353.2019.1667710.
Hajizadeh, S., A. Núñez, and D. M. J. Tax. 2016. “Semi-supervised rail defect detection from imbalanced image data.” IFAC-PapersOnLine 49 (3): 78–83. https://doi.org/10.1016/j.ifacol.2016.07.014.
Heidarysafa, M., K. Kowsari, L. Barnes, and D. Brown. 2019. “Analysis of railway accidents’ narratives using deep learning.” In Proc., 17th IEEE Int. Conf. on Machine Learning and Applications, 1446–1453. New York: IEEE.
Jamshidi, A., S. Faghih-Roohi, S. Hajizadeh, A. Núñez, R. Babuska, R. Dollevoet, Z. Li, and B. De Schutter. 2017. “A big data analysis approach for rail failure risk assessment.” Risk Anal. 37 (8): 1495–1507. https://doi.org/10.1111/risa.12836.
Lasisi, A., and N. Attoh-Okine. 2019. “An unsupervised learning framework for track quality index and safety.” Transp Infrastruct. Geotechnol. 7 (1). https://doi.org/10.1007/s40515-019-00087-6.
Li, H., D. Parikh, Q. He, B. Qian, Z. Li, D. Fang, and A. Hampapur. 2014. “Improving rail network velocity: A machine learning approach to predictive maintenance.” Transp. Res. Part C: Emerging Technol. 45 (Aug): 17–26. https://doi.org/10.1016/j.trc.2014.04.013.
Liu, X., V. L. Markine, H. Wang, and I. Y. Shevtsov. 2018. “Experimental tools for railway crossing condition monitoring (crossing condition monitoring tools).” Measurement 129 (Jul): 424–435. https://doi.org/10.1016/j.measurement.2018.07.062.
Lorente, A. G., D. F. Llorca, M. G. Velasco, J. A. R. Garcia, and F. S. Dominguez. 2014. “Detection of range-based rail gage and missing rail fasteners use of high-resolution two- and three-dimensional images.” Transp. Res. Rec. 2448 (1): 125–132. https://doi.org/10.3141/2448-15.
McGaughey, G., W. P. Walters, and B. Goldman. 2016. “Understanding covariate shift in model performance. F1000Research.” Chem. Inf. Sci. 5: 597. https://doi.org/10.12688/f1000research.8317.1.
PWayBlog. 2016. “Evaluation of the track quality.” Accessed January 11, 2021. https://pwayblog.com/2016/09/11/evaluation-of-track-quality/#TQI.
Qi, H., T. Xu, G. Wang, Y. Cheng, and C. Chen. 2020. “MYOLOv3-Tiny: A new convolutional neural network architecture for real-time detection of track fasteners.” Comput. Ind. 123 (Dec): 103303. https://doi.org/10.1016/j.compind.2020.103303.
Roghani, A., and M. T. Hendry. 2017. “Quantifying the impact of subgrade stiffness on track quality and the development of geometry defects.” J. Transp. Eng. 143 (7): 1–10. https://doi.org/10.1061/JTEPBS.0000043.
Sancho, L. C. B., J. A. P. Braga, and A. R. Andrade. 2021. “Optimizing maintenance decision in rails: A markov decision process approach.” J. Risk Uncertainty Eng. Syst. 7 (1): 04020051. https://doi.org/10.1061/AJRUA6.0001101.
Sharma, S., Y. Cui, Q. He, R. Mohammadi, and Z. Li. 2018. “Data-driven optimization of railway maintenance for track geometry.” Transp. Res. Part C: Emerging Technol. 90 (Sep): 34–58. https://doi.org/10.1016/j.trc.2018.02.019.
Shimodaira, H. 2000. “Improving predictive inference under covariate shift by weighting the log-likelihood function.” J. Stat. Plan. Inference 90 (2): 227–244. https://doi.org/10.1016/S0378-3758(00)00115-4.
Soleimanmeigouni, I., A. Ahmadi, and U. Kumar. 2018. “Track geometry degradation and maintenance modelling: A review.” Proc. Inst. Mech. Eng. 232 (1): 73–102. https://doi.org/10.1177/0954409716657849.
Song, Q., Y. Guo, J. Jiang, C. Liu, and M. Hu. 2019. High-speed railway fastener detection and localization method based on convolutional neural network. Lanham, MD: American Railway Engineering and Maintenance-of-Way Association.
Sugiyama, M., S. Nakajima, H. Kashima, P. V. Buenau, and M. Kawanabe. 2008. “Direct importance estimation with model selection and its application to covariate shift adaptation.” In Advances in Neural Information Processing Systems 20, 1433–1440. Cambridge, MA: MIT Press.
Zarembski, A. M., N. Attoh-Okine, D. Einbinder, H. Thompson, and T. Sussman. 2016. “How track geometry defects affect the development of rail defects.” In Proc., AREMA 2016 Annual Conf. & Exposition, 1081–1096. Maryland: American Railway Engineering and Maintenance-of-Way Association.
Information & Authors
Information
Published In
ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering
Volume 7 • Issue 3 • September 2021
Copyright
© 2021 American Society of Civil Engineers.
History
Received: Aug 19, 2020
Accepted: Feb 19, 2021
Published online: May 27, 2021
Published in print: Sep 1, 2021
Discussion open until: Oct 27, 2021
Authors
Metrics & Citations
Metrics
Citations
Download citation
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
Cited by
- Ana Ramos, Alexandre Castanheira-Pinto, Aires Colaço, Jesús Fernández-Ruiz, Pedro Alves Costa, Predicting Critical Speed of Railway Tracks Using Artificial Intelligence Algorithms, Vibration, 10.3390/vibration6040053, 6, 4, (895-916), (2023).
- Saeed Goodarzi, Hamed F. Kashani, Jimi Oke, Carlton L. Ho, Data-driven methods to predict track degradation: A case study, Construction and Building Materials, 10.1016/j.conbuildmat.2022.128166, 344, (128166), (2022).
- Ibrahim Balogun, Nii Attoh-Okine, Hybrid Reduction Techniques With Covariate Shift Optimization in High-Dimensional Track Geometry, Journal of Computing and Information Science in Engineering, 10.1115/1.4051597, 22, 1, (2021).
- Ibrahim Balogun, Mark Leadingham, Dominique Gulliot, Nii Attoh-Okine, Deep Learning Approach Towards Squat Isolation in a Multi-Embedded Track Geometry Defects, 2021 IEEE International Conference on Big Data (Big Data), 10.1109/BigData52589.2021.9671701, (4491-4495), (2021).