Abstract:
ITS needs accurate and timely traffic flow prediction. Graph-based neural networks have improved prediction results. Graph construction and model time complexity remain challenges. This paper proposes a multi-stream feature fusion approach to extract and integrate rich features from traffic data and construct graphs using a data-driven adjacent matrix instead of the distance-based matrix. The initial adjacent matrix is calculated from monitor station Spearman rank correlation coefficients and fine-tuned during training. We build a multi-stream feature fusion block (MFFB) module with a three-channel network and soft-attention mechanism. GCN, GRU, and FNN are three-channel networks that extract spatial, temporal, and other features. Integrating features uses soft-attention. A fully connected layer and convolutional layer predict with stacked MFFB modules. Our proposed approach outperforms state-of-the-art methods in two real-world traffic prediction tasks.
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
