Deep Stacked Bidirectional and Unidirectional LSTM Recurrent Neural Network for Network-wide Traffic Speed Prediction

Short-term traffic forecasting based on deep learning methods, especially long-term short memory (LSTM) neural networks, received much attention in recent years. However, the potential of deep learning methods is far from being fully exploited in terms of the depth of the architecture, the spatial scale of the prediction area, and the prediction power of spatial-temporal data. In this paper, a deep stacked bidirectional and unidirectional LSTM (SBU-LSTM) neural network is proposed, which considers both forward and backward dependencies of time series data, to predict the networkwide traffic speed. A bidirectional LSTM (BDLSM) layer is exploited to capture spatial features and bidirectional temporal dependencies from historical data. To the best of our knowledge, this is the first time that BDLSTM is applied as building blocks for a deep architecture model to measure the backward dependency of traffic data for prediction. A comparison with other classical and state-of-the-art models indicates that the proposed SBU-LSTM neural network achieves superior prediction performance for the whole traffic network in both accuracy and robustness.