tsbootstrap: Distribution-Free Uncertainty Quantification and Conformal Prediction for Time Series
📰 ArXiv cs.AI
Learn to quantify uncertainty in time series data using tsbootstrap, a library for distribution-free uncertainty quantification and conformal prediction
Action Steps
- Install tsbootstrap using pip to access its resampling and conformal prediction methods
- Use block, residual, sieve, or wild resampling to generate bootstrap samples from your time series data
- Apply adaptive conformal calibrators such as EnbPI, ACI, NexCP, or AgACI to quantify uncertainty in your predictions
- Evaluate the performance of different resampling and calibration methods using metrics such as coverage probability and interval width
- Integrate tsbootstrap into your existing time series forecasting pipeline to improve uncertainty quantification and conformal prediction capabilities
Who Needs to Know This
Data scientists and machine learning engineers working with time series data can benefit from using tsbootstrap to improve the accuracy and reliability of their predictions
Key Insight
💡 tsbootstrap provides a flexible and typed API for resampling and conformal prediction, allowing users to easily switch between different methods and evaluate their performance
Share This
📊🕰️ Introducing tsbootstrap: a library for distribution-free uncertainty quantification and conformal prediction in time series data #timeseries #uncertaintyquantification #conformalprediction
Key Takeaways
Learn to quantify uncertainty in time series data using tsbootstrap, a library for distribution-free uncertainty quantification and conformal prediction
Full Article
Title: tsbootstrap: Distribution-Free Uncertainty Quantification and Conformal Prediction for Time Series
Abstract:
arXiv:2607.06690v1 Announce Type: cross Abstract: Finance, sensing, and demand streams violate the exchangeability that IID conformal prediction and the IID bootstrap assume, and existing libraries implement either a general resampling engine or conformal calibration without the other. tsbootstrap provides block, residual, sieve, and wild resampling, classical bootstrap confidence intervals, and adaptive conformal calibrators (EnbPI, ACI, NexCP, AgACI) through a single typed API in which a speci
Abstract:
arXiv:2607.06690v1 Announce Type: cross Abstract: Finance, sensing, and demand streams violate the exchangeability that IID conformal prediction and the IID bootstrap assume, and existing libraries implement either a general resampling engine or conformal calibration without the other. tsbootstrap provides block, residual, sieve, and wild resampling, classical bootstrap confidence intervals, and adaptive conformal calibrators (EnbPI, ACI, NexCP, AgACI) through a single typed API in which a speci
DeepCamp AI