Accurately predicting the bus travel time is important for improving bus service levels and enhancing the attractiveness of the bus service. This study first integrated multi-source data to construct multiple influencing factors, including bus arrival time data, spatial attributes of bus stops and routes, smart card data, and weather data. Random forest was then used to quantitatively analyze the impact of various influencing factors on bus travel time. The convolutional neural network-long short-term memory-attention mechanism (CNN-LSTM-ATTENTION) model was proposed to predict the bus travel time. One month’s multi-source data from Beijing was used to build the model. The results indicate that the travel time of the previous bus has the highest importance, therefore it is necessary to incorporate this feature for predicting bus arrival time to enhance prediction accuracy. The CNN-LSTM-ATTENTION model can reduce the mean square error by up to 29.7% compared to the traditional LightGBM model. Considering the travel time of the previous bus reduces the mean square error by 12.8% compared to without considering it.
Bus Travel Time Prediction Based on Multiple Influencing Factors and CNN-LSTM-ATTENTION Network
Lect. Notes Electrical Eng.
International Conference on SmartRail, Traffic and Transportation Engineering ; 2024 ; Chongqing, China October 25, 2024 - October 27, 2024
Advances and Applications in SmartRail, Traffic, and Transportation Engineering ; Chapter : 20 ; 205-213
2025-07-19
9 pages
Article/Chapter (Book)
Electronic Resource
English
Travel time prediction with LSTM neural network
IEEE | 2016
|Travel Time Probability Prediction Based on Constrained LSTM Quantile Regression
DOAJ | 2023
|Bus travel time prediction method based on space-time diagram attention network
European Patent Office | 2025
|