Abstract:
Real-time traffic demand can be revealed by accurate metro ridership forecasting. Deep learning algorithms are widely used to capture spatio-temporal features in metro riding behavior due to their superior performance. Current deep learning models use regular convolutional operations, which can barely provide satisfactory accuracy due to either a lack of realistic traffic network topology or insufficient spatiotemporal patterns. This study proposes a parallel-structured deep learning model with a Graph Convolution Network and a stacked Bidirectional unidirectional Long short-term Memory network (GCN-SBULSTM) to improve metro ridership prediction. The GCN module uses a K-hop matrix to capture the dynamic spatial correlation among metro stations. SBULSTM can learn complex temporal features with stacked recurrent layers and considers ridership time series backward and forward. The model is tested on three real-world metro ridership datasets. GCN-SBULSTM outperforms state-of-the-art prediction models and greatly improves training efficiency.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
