The Journal of The Korea Institute of Intelligent Transport Systems (한국ITS학회 논문지)

The Korea Institute of Intelligent Transport Systems (한국ITS학회)
권호 표지
DOI QR코드

DOI QR Code

Development of Highway Traffic Information Prediction Models Using the Stacking Ensemble Technique Based on Cross-validation

스태킹 앙상블 기법을 활용한 고속도로 교통정보 예측모델 개발 및 교차검증에 따른 성능 비교

  • Yoseph Lee (Dept. of Transportation Eng., Ajou University) ;
  • Seok Jin Oh (Dept. of Civil and Environmental Eng., Univ. of Honam) ;
  • Yejin Kim (Dept. of Transportation Eng., Ajou University) ;
  • Sung-ho Park (Dept. of Transportation Eng., Ajou University) ;
  • Ilsoo Yun (Dept. of Transportation Eng., Ajou University)
  • 이요셉 (아주대학교 교통공학과) ;
  • 오석진 (호남대학교 토목환경공학과) ;
  • 김예진 (아주대학교 교통공학과) ;
  • 박성호 (아주대학교 교통연구센터) ;
  • 윤일수 (아주대학교 교통시스템공학과)
  • Received : 2023.10.03
  • Accepted : 2023.11.14
  • Published : 2023.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Accurate traffic information prediction is considered to be one of the most important aspects of intelligent transport systems(ITS), as it can be used to guide users of transportation facilities to avoid congested routes. Various deep learning models have been developed for accurate traffic prediction. Recently, ensemble techniques have been utilized to combine the strengths and weaknesses of various models in various ways to improve prediction accuracy and stability. Therefore, in this study, we developed and evaluated a traffic information prediction model using various deep learning models, and evaluated the performance of the developed deep learning models as a stacking ensemble. The individual models showed error rates within 10% for traffic volume prediction and 3% for speed prediction. The ensemble model showed higher accuracy compared to other models when no cross-validation was performed, and when cross-validation was performed, it showed a uniform error rate in long-term forecasting.

정확도가 높은 교통정보 예측은 지능형교통체계(intelligent transport systems, ITS)를 통한 교통 시설 이용자들의 혼잡 경로 회피 안내 등에서 활용되는 중요한 기능이다. 정확한 교통정보예측을 위해 다양한 딥러닝 모델들이 발전되어 왔다. 최근에는 앙상블 기법을 활용하여 다양한 모델들의 장단점을 결합하여 예측 정확도와 안정성을 높이고 있다. 따라서, 본 연구에서는 다양한 딥러닝 모델들을 활용하여 교통정보 예측 모델을 개발하였으며, 개발된 딥러닝 모델들을 스태킹 앙상블(stacking ensemble)하여 성능을 개선하였다. 개별 모델들은 교통량 예측에서 10% 이내의 오차율을, 속도 예측에서 3% 이내의 오차율을 보였다. 앙상블 모델은 교차검증을 수행하지 않았을 때, 타 모델과 비교하여 더욱 높은 정확도를 보였다. 교차검증을 수행한 앙상블 모델은 장기예측에서 타 모델보다 균일한 오차율을 보이는 것으로 나타났다.

Keywords

Acknowledgement

본 논문은 국토교통부 자율주행 기술개발 혁신사업 '주행 및 충돌상황 대응 안전성 평가기술개발(RS-2021-KA160637)' 과제 지원에 의해 수행되었습니다.

References

  1. Al-Hajj, R., Assi, A. and Fouad, M.(2019), "Stacking-based ensemble of support vector regressors for one-day ahead solar irradiance prediction", 8th International Conference on Renewable Energy Research and Applications(ICRERA), pp.428-433.
  2. Asencio-Cortes, G., Florido, E., Troncoso, A. and Martinez-Alvarez, F.(2016), "A novel methodology to predict urban traffic congestion with ensemble learning", Soft Computing, vol. 11, no. 20, pp.4205-4216.
  3. Brownlee, J.(2018), "Better Deep Learning: Train Faster, Reduce Overfitting, and Make better Predictions", Machine Learning Mastery, p.540.
  4. Chung, J., Gulcehre, C., Cho, K. and Bengio, Y.(2014), Empirical evaluation of gated recurrent neural networks on sequence modeling, arXiv preprint, arXiv:1412.3555.
  5. Creamer, G. G. and Freund, Y.(2010), "Automated Trading with Boosting and Expert Weighting", Quantitative Finance, vol. 3, pp.147-163. https://doi.org/10.1080/14697680903104113
  6. Del Ser, J., Lana, I., Bilbao, M. N. and Vlahogianni, E. I.(2019), "Road traffic forecasting using stacking ensembles of echo state networks", 2019 IEEE Intelligent Transportation Systems Conference(ITSC), pp.2591-2597.
  7. Eom, H., Kim, J. and Choi, S. O.(2020), "Machine learning-based corporate default risk prediction model verification and policy recommendation: Focused on improvement through stacking ensemble model", Journal of Intelligence and Information Systems, vol. 26, no. 2, pp.105-129. https://doi.org/10.13088/JIIS.2020.26.2.105
  8. Fink, W. G.(1995), "Intelligent transportation systems", IEEE 1995 Microwave and Millimeter-Wave, Monolithic Circuits Symposium, Digest of Papers.
  9. Hwang, D. and Jo, K.(2021), "Prediction of odor concentration from pig production based on machine learning: A case study of hydrogen sulfide", Journal of Odor and Indoor Environments, vol. 20, no. 4, pp.311-321. https://doi.org/10.15250/joie.2021.20.4.311
  10. Jang, P. S.(2022), "Performance Analysis of Trading Strategy using Gradient Boosting Machine Learning and Genetic Algorithm", Journal of the Korea society of Computer and Information, vol. 27, no. 11, pp.147-155. https://doi.org/10.9708/JKSCI.2022.27.11.147
  11. Jeon, B., Kang, J. and Chung, Y.(2021), "AutoML and CNN-Based Soft-Voting Ensemble Classification Model For Road Traffic Emerging Risk Detection", Journal of Convergence for Information Technology, vol. 11, no. 7, pp.14-20. https://doi.org/10.22156/CS4SMB.2021.11.07.014
  12. Kang, H. and Noh, M.(2022), "Classifying the security of pedestrian accidents using ensemble machine learning algorithms: A case study of Daejeon City", Journal of Digital Convergence, vol. 20, no. 5, pp.39-46. https://doi.org/10.14400/JDC.2022.20.5.039
  13. Kim, D. and Lee, K.(2020a), "Forecasting the Container Volumes of Busan Port using LSTM", Journal of Korea Port Economic Association, vol. 36, no. 2, pp.53-62. https://doi.org/10.38121/kpea.2020.06.36.2.53
  14. Kim, H., Lee, S. and Hwang, J.(2023), "Development of a Speed Prediction Model for URban Network Based on Gated Recurrent Unit", The Journal of the Korea Institute of Intelligent Transport Systems, vol. 22, no. 1, pp.103-114. https://doi.org/10.12815/kits.2023.22.1.103
  15. Kim, Y. and Park, H.(2021), "Modeling and Predicting South Korea's Daily Electric Demand Using DNN and LSTM", Journal of Climate Change Research, vol. 12, no. 3, pp.241-253. https://doi.org/10.15531/KSCCR.2021.12.3.241
  16. Kim, Y., Kim, J., Han, Y., Kim, J. and Hwang, J.(2020b), "Development of Traffic Speed Prediction Model Reflecting Spatio-temporal Impact based on Deep Neural Network", The Journal of the Korea Institute of Intelligent Transport System, vol. 19, no. 1, pp.1-16. https://doi.org/10.12815/kits.2020.19.1.1
  17. Ko, S. J., Yun, H. Y. and Shin, D. M.(2018), "Electronic Demand Data Prediction using Bidirectional Long Short Term Memory Networks", Journal of Software Assessment and Valuation, vol. 14, no. 1, pp.33-40.
  18. Lee, H., Cho, J., Kang, H. and Hwang, J.(2019), "Applicability study on urban flooding risk criteria estimation algorithm using cross-validation and SVM", Journal of Korea Water Resources Association, vol. 52, no. 12, pp.963-973. https://doi.org/10.3741/JKWRA.2019.52.12.963
  19. Lee, Y., Jin, H., Kim, Y., Park, S. and Yun, I.(2023), "Methodology for Developing a Predictive Model for Highway Traffic Information Using LSTM", The Journal of the Korea Institute of Intelligent Transport System, vol. 22, no. 5, pp.01-18. https://doi.org/10.12815/kits.2023.22.5.1
  20. Liu, J., Wu, N., Qiao, Y. and Li, Z.(2020), "Short-term traffic flow forecasting using ensemble approach based on deep belief networks", IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 1, pp.404-417. https://doi.org/10.1109/TITS.2020.3011700
  21. Park, H.(2022), "Deep Learning Model for Traffic Speed Prediction Considering Spatial-Temporal Correlation", The Journal of Korean Institute of Information Technology, vol. 20, no. 3, pp.19-26.
  22. Rokach, L.(2010), "Ensemble-based classifiers", Artificial Intelligence Review, vol. 33, no. 1, pp.1-39. https://doi.org/10.1007/s10462-009-9124-7
  23. Ruder, S.(2017), An overview of gradient descent optimization algorithms, arXiv preprint arXiv:1609.04747.
  24. Seo, H.(2022), "Development of public bicvle demand forecasting model using cross-validation based staking methods", The Statistical Review, vol. 22, pp.31-38.
  25. Shi, J., Jain, M. and Narasimhan, G.(2022), Time series forecasting (tsf) using various deep learning models, arXiv preprint arXiv:2204.11115.
  26. Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I. and Salakhutdinov, R.(2014), "Dropout: A simple way to prevent neural networks from overfitting", The Journal of Machine Learning Research, vol. 15, no. 1, pp.1929-1958.
  27. Yang, S., Wu, J., Du, Y., He, Y. and Chen, X.(2017), "Ensemble learning for short-term traffic prediction based on gradient boosting machine", Journal of Sensors, pp.1-15.
  28. Zheng, G., Chai, W. K. and Katos, V.(2019), "An ensemble model for short-term traffic prediction in smart city transportation system", 2019 IEEE Global Communications Conference (GLOBECOM), pp.1-6.
  29. Zheng, G., Chai, W. K., Katos, V. and Walton, M.(2021), "A joint temporal-spatial ensemble model for short-term traffic prediction", Neurocomputing, vol. 457, pp.26-39. https://doi.org/10.1016/j.neucom.2021.06.028
  30. Zhong, G., Yin, T., Li, L., Zhang, J., Zhang, H. and Ran, B.(2020), "Bus travel time prediction based on ensemble learning methods", IEEE Intelligent Transportation Systems Magazine, vol. 14, no. 2, pp.174-189.