Conformalized Quantile Regression

Abstract

Conformal prediction is a technique for constructing prediction intervals that attain valid coverage in finite samples, without making distributional assumptions. Despite this appeal, existing conformal methods can be unnecessarily conservative because they form intervals of constant or weakly varying length across the input space. In this paper we propose a new method that is fully adaptive to heteroscedasticity. It combines conformal prediction with classical quantile regression, inheriting the advantages of both. We establish a theoretical guarantee of valid coverage, supplemented by extensive experiments on popular regression datasets. We compare the efficiency of conformalized quantile regression to other conformal methods, showing that our method tends to produce shorter intervals.

19 pages, 8 figures, 1 table

References in corpus (6)

• Adam: A Method for Stochastic Optimization
• Estimating conditional quantiles with the help of the pinball loss
• Regression Conformal Prediction with Nearest Neighbours
• High-Quality Prediction Intervals for Deep Learning: A Distribution-Free, Ensembled Approach
• Reliable Prediction Intervals with Regression Neural Networks
• Trimmed Conformal Prediction for High-Dimensional Models

Cited by in corpus (21)

• Testing for Outliers with Conformal p-values
• A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantification
• Neural Networks for Extreme Quantile Regression with an Application to Forecasting of Flood Risk
• Flexible distribution-free conditional predictive bands using density estimators
• White Paper Machine Learning in Certified Systems
• CD-split and HPD-split: efficient conformal regions in high dimensions
• Differentiable Ranks and Sorting using Optimal Transport
• Exchangeability, Conformal Prediction, and Rank Tests
• With Malice Towards None: Assessing Uncertainty via Equalized Coverage
• ML-AQP: Query-Driven Approximate Query Processing based on Machine Learning
• Conformal Anomaly Detection on Spatio-Temporal Observations with Missing Data
• Model-free Bootstrap and Conformal Prediction in Regression: Conditionality, Conjecture Testing, and Pertinent Prediction Intervals
• Conformal inference for random objects
• Understanding the Under-Coverage Bias in Uncertainty Estimation
• Efficient and minimal length parametric conformal prediction regions
• Residual permutation test for regression coefficient testing
• Distributional Robust Kelly Gambling: Optimal Strategy under Uncertainty in the Long-Run
• T-SCI: A Two-Stage Conformal Inference Algorithm with Guaranteed Coverage for Cox-MLP
• AutoCP: Automated Pipelines for Accurate Prediction Intervals
• Conformal Bayesian Computation
• Synthesizing Machine Learning Programs with PAC Guarantees via Statistical Sketching