Open AccessForecasting the railway freight volume in China based on combined PSO-LSTM model
Author(s) -
Yue-Ying Qiu,
Qiong Zhang,
Ming Lei
Publication year - 2020
Publication title -
journal of physics: conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
ISSN - 1742-6588
DOI - 10.1088/1742-6596/1651/1/012029
Subject(s) - volume (thermodynamics) , computer science , artificial neural network , selection (genetic algorithm) , traffic volume , artificial intelligence , engineering , transport engineering , quantum mechanics , physics
Considering the shortcomings of the current railway freight volume prediction model, this paper proposes a railway freight volume forecasting model based on PSO-LSTM. LSTM can learn long-span nonlinear time series effectively, and it is highly efficient in training forecasting of railway freight volume data. In order to improve the accuracy of parameter selection, PSO is used to optimize neural network model. The simulation results show that the PSO-LSTM model proposed has lower prediction error and higher prediction accuracy than the traditional LSTM model GA-LSTM model. This indicates that this model is an effective prediction model for railway freight volume.