Research map / Time series / Causal and generative time series
Split Conformal Prediction and Non-Exchangeable Data
Roberto I. Oliveira, Paulo Orenstein, Thiago Ramos, João Vitor Romano
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| use plain split CP instead of NexCP weighting | NexCP · Spatiotemporal temperature | interval size (deg C): 9.605 vs 9.289 | No clear effect |
| use plain split CP instead of DtACI | DtACI · Spatiotemporal temperature | average time (min): 75.15 -> 2.85 | Improved |
| use plain split CP instead of NexCP | NexCP · Spatiotemporal temperature | average time (min): 1884.92 -> 2.85 | Improved |
| add dynamically tuned adaptive conformal inference (DtACI) | Split CP · Spatiotemporal temperature | coverage SD over space: 0.130 -> 0.096 | Improved |
| add dynamically tuned adaptive conformal inference (DtACI) | Split CP · Spatiotemporal temperature | interval size (deg C): 9.605 -> 9.658 | Hurt |
| decrease NexCP decay parameter rho from 0.99 to 0.1 (ablation) | NexCP · Spatiotemporal temperature | average coverage: 90.9% -> 89.5% | Hurt |
| use split CP instead of EnbPI | EnbPI · Cycle-graph random walk | coverage: 86.6% -> 91.4% | Improved |
| use split CP instead of EnbPI | EnbPI · Cycle-graph random walk | interval size: 0.15 -> 0.34 | Hurt |
| increase calibration set size from 500 to 5000 (ablation) | Split CP · Financial eurusd high volatility | conditional coverage: 87.64% -> 89.85% | Improved |
| increase calibration set size from 500 to 5000 (ablation) | Split CP · Financial bcousd high volatility | conditional coverage: 87.10% -> 89.33% | Improved |
| increase autoregressive coefficient (stronger dependence) (ablation) | Split CP · AR(1) simulation | marginal coverage | No clear effect |
Closest papers by meaning
- Non-exchangeable Conformal Prediction with Optimal Transport: Tackling Distribution Shift with Unlabeled Data (NeurIPS 2025 · similarity 0.91)
- Non-Exchangeable Conformal Risk Control (ICLR 2024 · similarity 0.91)
- Conformal Predictions under Markovian Data (ICML 2024 · similarity 0.90)
- Flow-Based Conformal Predictive Distributions (NeurIPS 2026 · similarity 0.88)
- Conformal prediction for multi-dimensional time series by ellipsoidal sets (ICML 2024 · similarity 0.88)
- Improving the Statistical Efficiency of Cross-Conformal Prediction (ICML 2025 · similarity 0.88)
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.
