Research map / Video / Video segmentation and tracking
Deep Multi-View Spatial-Temporal Network for Taxi Demand Prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Combine local CNN spatial features, LSTM temporal modeling, and semantic region embeddings | Taxi demand prediction · Guangzhou taxi demand test set | MAPE: 0.1971 -> 0.1616 | Helped |
| Add semantic view to temporal view (ablation) | DMVST-Net · Guangzhou taxi demand test set | MAPE: 0.1721 -> 0.1708 | Helped |
| Add neighboring-region demand features (ablation) | Temporal view · Guangzhou taxi demand test set | MAPE: 0.1721 -> 0.1710 | Helped |
| Use a local CNN for spatial features instead of simple neighboring-region inputs (ablation) | Temporal + spatial model · Guangzhou taxi demand test set | MAPE: 0.1710 -> 0.1640 | Helped |
| Add local CNN spatial view to temporal view (ablation) | DMVST-Net · Guangzhou taxi demand test set | MAPE: 0.1721 -> 0.1640 | Helped |
| Use only the last demand value as LSTM input, without context features (ablation) | LSTM temporal view · Guangzhou taxi demand test set | MAPE | Hurt |
| Increase LSTM sequence length up to 4 hours (ablation) | DMVST-Net · Guangzhou taxi demand test set | MAPE | Helped |
| Increase LSTM sequence length beyond 4 hours (ablation) | DMVST-Net · Guangzhou taxi demand test set | MAPE | Hurt |
| Increase local CNN neighborhood size from 9×9 to 13×13 (ablation) | DMVST-Net · Guangzhou taxi demand test set | MAPE: more than 3.46% | Hurt |
| Reduce local CNN neighborhood size from 9×9 to 5×5 (ablation) | DMVST-Net · Guangzhou taxi demand test set | MAPE | Hurt |
| Add spatial and temporal modeling rather than relying on context-feature regression | Taxi demand prediction · Guangzhou taxi demand test set | MAPE: 0.1840 -> 0.1616 | Helped |
| Use the combined multi-view model rather than temporal view alone (ablation) | DMVST-Net · Guangzhou taxi demand test set | RMSE: 9.812 -> 9.642 | Helped |
Closest papers by meaning
- Spatiotemporal Multi-Graph Convolution Network for Ride-Hailing Demand Forecasting (AAAI 2019 · similarity 0.90)
- Coupled Layer-wise Graph Convolution for Transportation Demand Prediction (AAAI 2021 · similarity 0.89)
- Revisiting Spatial-Temporal Similarity: A Deep Learning Framework for Traffic Prediction (AAAI 2019 · similarity 0.89)
- Learning Heterogeneous Spatial-Temporal Representation for Bike-Sharing Demand Prediction (AAAI 2019 · similarity 0.88)
- Community-Aware Multi-Task Transportation Demand Prediction (AAAI 2021 · similarity 0.87)
- Dynamic Spatial-Temporal Graph Convolutional Neural Networks for Traffic Forecasting (AAAI 2019 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.