Running trains on time is the biggest challenge of the railway industry, so the prediction of train delay can help to better train movement and more precise calculations of a train’s running condition, which helps railway departments to make decisions for other trains and convey the same message to commuters. There are many reasons for train delays, most of them are commutator traffic, maintenance, faults, worse weather, and train movement strategies. Train delay basically depends on departure time, in conventional methods to measure delay consider only departure time. But departure time is not only one factor that decides train delay there are two more factors which cause delay, i.e., distance/space and time between two trains. In this study, we introduce a machine learning framework to predict train delay, called the train spatiotemporal graph convolutional network (TSTGCN) Zhang (Trans Intell Transp Syst 23:2434–2444, 2021). This model can help to predicting delay time of one train and also predicts the total number of arrival delays caused by all trains Jing (Research on delay prediction of high speed railway train based on data analysis. Chengdu, 2019). Model required certain data for prediction; these are based on data collected recent, daily, and weekly; and each consisting of a spatiotemporal attention mechanism and spatiotemporal convolution to effectively capture spatiotemporal features Jing (Research on delay prediction of high speed railway train based on data analysis. Chengdu, 2019). Combination of three components produces the final result. If we compare TSTGCN with existing advanced baselines in predicting train delays, we can clearly see TSTGCN surpasses other.
Train Delay Prediction Using Machine Learning
Algorithms for Intelligent Systems
International Conference on Paradigms of Communication, Computing and Data Analytics ; 2023 ; Delhi, India April 22, 2023 - April 23, 2023
Proceedings of International Conference on Paradigms of Communication, Computing and Data Analytics ; Kapitel : 54 ; 663-674
11.10.2023
12 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Enhancing freight train delay prediction with simulation‐assisted machine learning
DOAJ | 2024
|Enhancing freight train delay prediction with simulation‐assisted machine learning
Wiley | 2024
|Prediction of Train Delay System in Indian Railways Using Machine Learning Techniques: Survey
Springer Verlag | 2022
|Flight Delay Prediction using Machine Learning Model
IEEE | 2022
|