Scholarlink

In order to alleviate urban traffic congestion and bolster the efficiency of urban traffic systems, precise forecasting of traffic flow across diverse areas is of paramount importance. To facilitate short-term forecasting of urban traffic patterns, a novel deep neural network architecture, termed the CBA-Model, underpinned by an encoder-decoder framework, has been introduced. LSTM, GRU and SAEs are introduced in the encoder and decoder structure, and the attention mechanism is integrated into the architecture to augment the emphasis on pivotal features. Then three traffic flow prediction models are constructed and the performance of the models is assessed through two survey datasets. The empirical findings indicate that the novel model, equipped with an LSTM-based encoder, has yielded an MAE score of 7.72 and an RMSE figure of 11.26, thereby affirming its exceptional predictive prowess.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A New Short-Term Traffic Flow Prediction Method Based on CBA-Model Architecture

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Mingkaiqi Wang,
  • Dizhi Guo,
  • Mengxin Li,
  • Chao Huang

摘要

In order to alleviate urban traffic congestion and bolster the efficiency of urban traffic systems, precise forecasting of traffic flow across diverse areas is of paramount importance. To facilitate short-term forecasting of urban traffic patterns, a novel deep neural network architecture, termed the CBA-Model, underpinned by an encoder-decoder framework, has been introduced. LSTM, GRU and SAEs are introduced in the encoder and decoder structure, and the attention mechanism is integrated into the architecture to augment the emphasis on pivotal features. Then three traffic flow prediction models are constructed and the performance of the models is assessed through two survey datasets. The empirical findings indicate that the novel model, equipped with an LSTM-based encoder, has yielded an MAE score of 7.72 and an RMSE figure of 11.26, thereby affirming its exceptional predictive prowess.