TrafficBERT: Pre-trained model with large-scale data for long-range traffic flow forecasting

作者：

Highlights：

• Proposing a deep learning model to predict long-range traffic flow forecasting.

• TrafficBERT is a modifies of the BERT structure for traffic flow forecasting.

• Increasing the generality of the model by pre-training data on various roads.

摘要

•Proposing a deep learning model to predict long-range traffic flow forecasting.•TrafficBERT is a modifies of the BERT structure for traffic flow forecasting.•Increasing the generality of the model by pre-training data on various roads.

论文关键词：Traffic flow,Big data,Pre-trained model,BERT

论文评审过程：Received 12 July 2020, Revised 6 June 2021, Accepted 5 August 2021, Available online 22 August 2021, Version of Record 26 August 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115738