Online prediction of arrival and departure times in each station for passenger trains using machine learning methods
The prediction of delays and their reduction in all modes of passenger transportation, especially rail transportation, is of great importance and annually attracts the attention of many researchers. Train delays can be anticipated by predicting the arrival times of trains at stations. In this paper,...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-06-01
|
| Series: | Transportation Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666691X24000253 |
| _version_ | 1797214826311712768 |
|---|---|
| author | Shekoofeh Vafaei Masoud Yaghini |
| author_facet | Shekoofeh Vafaei Masoud Yaghini |
| author_sort | Shekoofeh Vafaei |
| collection | DOAJ |
| description | The prediction of delays and their reduction in all modes of passenger transportation, especially rail transportation, is of great importance and annually attracts the attention of many researchers. Train delays can be anticipated by predicting the arrival times of trains at stations. In this paper, a train operated by Raja Company, which travels daily on the Tehran-Mashhad route, has been investigated. This train route consists of 50 stations, of which five main stations, including Tehran, Garmsar, Semnan, Shahrud, and Mashhad, have been selected to predict the train's arrival and departure times at each of these stations. For this purpose, data related to the train timetable and the operations carried out at these five main stations over three years from 2018 to 2020 have been collected. Then, modeling was conducted to predict real-time arrival and departure times for each of these stations. Artificial neural networks, random forest regression, gradient boosting regression, and extreme gradient boosting regression were used for prediction modeling. After evaluating these models, the approach that yielded the best results based on the experimental data was selected for predicting the arrival and departure times at each station. |
| first_indexed | 2024-04-24T11:20:20Z |
| format | Article |
| id | doaj.art-8663f40ed3d743fbbfc667c49be2076c |
| institution | Directory Open Access Journal |
| issn | 2666-691X |
| language | English |
| last_indexed | 2024-04-24T11:20:20Z |
| publishDate | 2024-06-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Transportation Engineering |
| spelling | doaj.art-8663f40ed3d743fbbfc667c49be2076c2024-04-11T04:42:00ZengElsevierTransportation Engineering2666-691X2024-06-0116100250Online prediction of arrival and departure times in each station for passenger trains using machine learning methodsShekoofeh Vafaei0Masoud Yaghini1Department of Railway Engineering, Iran University of Science and Technology, Tehran, IranCorresponding author.; Department of Railway Engineering, Iran University of Science and Technology, Tehran, IranThe prediction of delays and their reduction in all modes of passenger transportation, especially rail transportation, is of great importance and annually attracts the attention of many researchers. Train delays can be anticipated by predicting the arrival times of trains at stations. In this paper, a train operated by Raja Company, which travels daily on the Tehran-Mashhad route, has been investigated. This train route consists of 50 stations, of which five main stations, including Tehran, Garmsar, Semnan, Shahrud, and Mashhad, have been selected to predict the train's arrival and departure times at each of these stations. For this purpose, data related to the train timetable and the operations carried out at these five main stations over three years from 2018 to 2020 have been collected. Then, modeling was conducted to predict real-time arrival and departure times for each of these stations. Artificial neural networks, random forest regression, gradient boosting regression, and extreme gradient boosting regression were used for prediction modeling. After evaluating these models, the approach that yielded the best results based on the experimental data was selected for predicting the arrival and departure times at each station.http://www.sciencedirect.com/science/article/pii/S2666691X24000253Passenger trainArrival timeDeparture timeMachine learningOnline predictionTransportation network |
| spellingShingle | Shekoofeh Vafaei Masoud Yaghini Online prediction of arrival and departure times in each station for passenger trains using machine learning methods Transportation Engineering Passenger train Arrival time Departure time Machine learning Online prediction Transportation network |
| title | Online prediction of arrival and departure times in each station for passenger trains using machine learning methods |
| title_full | Online prediction of arrival and departure times in each station for passenger trains using machine learning methods |
| title_fullStr | Online prediction of arrival and departure times in each station for passenger trains using machine learning methods |
| title_full_unstemmed | Online prediction of arrival and departure times in each station for passenger trains using machine learning methods |
| title_short | Online prediction of arrival and departure times in each station for passenger trains using machine learning methods |
| title_sort | online prediction of arrival and departure times in each station for passenger trains using machine learning methods |
| topic | Passenger train Arrival time Departure time Machine learning Online prediction Transportation network |
| url | http://www.sciencedirect.com/science/article/pii/S2666691X24000253 |
| work_keys_str_mv | AT shekoofehvafaei onlinepredictionofarrivalanddeparturetimesineachstationforpassengertrainsusingmachinelearningmethods AT masoudyaghini onlinepredictionofarrivalanddeparturetimesineachstationforpassengertrainsusingmachinelearningmethods |