Research map / Time series / Causal and generative time series
RiskOracle: A Minute-Level Citywide Traffic Accident Forecasting Framework
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 |
|---|---|---|---|
| Add PKDE to transform zero labels into risk-informed negative values (ablation) | RiskOracle · NYC 30-minute and 10-minute | Acc@20 / Acc@6: 48% -> 70%; 28% -> 45% | Helped |
| Add ST-DFM sparse-sensing co-sensing (ablation) | RiskOracle · NYC 30-minute and 10-minute | Acc@20 / Acc@6: 69% -> 70%; 43% -> 45% | Helped |
| Add time-varying overall affinity (ablation) | RiskOracle · NYC 30-minute and 10-minute | Acc@20 / Acc@6: 34% -> 70%; 29% -> 45% | Helped |
| Add the differential feature generator (ablation) | RiskOracle · NYC 30-minute and 10-minute | Acc@20 / Acc@6: 63% -> 70%; 42% -> 45% | Helped |
| Add multi-task learning with HARS (ablation) | RiskOracle · NYC 30-minute and 10-minute | Acc@20 / Acc@6: 65% -> 70%; 40% -> 45% | Helped |
| Use RiskOracle rather than ST-ResNet | NYC · NYC 30-minute | MSE: 0.0627 -> 0.1085 | Hurt |
| Use RiskOracle rather than the best baseline | NYC · NYC 30-minute | Acc@20: 48.04% -> 70.53% | Helped |
| Use RiskOracle rather than SDCAE | NYC · NYC 10-minute | MSE: 0.0076 -> 0.0452 | Hurt |
| Use RiskOracle rather than the best baseline | NYC · NYC 10-minute | Acc@6: 24.53% -> 45.18% | Helped |
| Use RiskOracle rather than SDCAE | SIP · SIP 30-minute | MSE: 3.210 -> 3.270 | Hurt |
| Use RiskOracle rather than ST-ResNet | SIP · SIP 30-minute | Acc@5: 60.73% -> 63.15% | Helped |
| Use RiskOracle rather than ST-ResNet | SIP · SIP 10-minute | MSE: 3.180 -> 3.029 | Helped |
Closest papers by meaning
- TraffiDent: A Dataset for Understanding the Interplay Between Traffic Dynamics and Incidents (NeurIPS 2025 · similarity 0.86)
- Accident Anticipation via Temporal Occurrence Prediction (NeurIPS 2025 · similarity 0.86)
- FT-AED: Benchmark Dataset for Early Freeway Traffic Anomalous Event Detection (NeurIPS 2024 · similarity 0.83)
- Progressive Risk Estimation for Accident Anticipation (NeurIPS 2026 · similarity 0.83)
- Spatiotemporal-aware Trend-Seasonality Decomposition Network for Traffic Flow Forecasting (AAAI 2025 · similarity 0.82)
- Harnessing Event Sensory Data for Error Pattern Prediction in Vehicles: A Language Model Approach (AAAI 2025 · similarity 0.82)
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.