Conditional validity of inductive conformal predictors

DOI10.1007/S10994-013-5355-6zbMATH Open1273.68307DBLPjournals/ml/Vovk13arXiv1209.2673OpenAlexW2166693382WikidataQ62046668 ScholiaQ62046668MaRDI QIDQ374136FDOQ374136

Authors: Vladimir Vovk 

Publication date: 22 October 2013

Published in: Machine Learning (Search for Journal in Brave)

Abstract: Conformal predictors are set predictors that are automatically valid in the sense of having coverage probability equal to or exceeding a given confidence level. Inductive conformal predictors are a computationally efficient version of conformal predictors satisfying the same property of validity. However, inductive conformal predictors have been only known to control unconditional coverage probability. This paper explores various versions of conditional validity and various ways to achieve them using inductive conformal predictors and their modifications.

Full work available at URL: https://arxiv.org/abs/1209.2673

Recommendations

• Conformal prediction beyond exchangeability
• scientific article; zbMATH DE number 6306019
• A conditional property of invariant confidence and prediction regions
• Combination of inductive Mondrian conformal predictors
• Conditional predictive inference for stable algorithms
• Conformal prediction: a unified review of theory and new challenges
• Conformal Prediction: A Gentle Introduction
• Conformal prediction under hypergraphical models
• Conditional properties of prediction intervals

zbMATH Keywords

ROC curvesboostingMARTbatch mode of learningconditional validityinductive conformal predictorsspam detection

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Learning and adaptive systems in artificial intelligence (68T05)

Cites Work

• A decision-theoretic generalization of on-line learning and an application to boosting
• Greedy function approximation: A gradient boosting machine.
• Introduction to nonparametric estimation
• Title not available (Why is that?)
• Title not available (Why is that?)
• Distribution-Free Prediction Sets
• Distribution-free Prediction Bands for Non-parametric Regression
• Stochastic gradient boosting.
• Title not available (Why is that?)
• Tutorial on practical prediction theory for classification
• Data analysis and graphics using R -- an example-based approach.
• Title not available (Why is that?)
• Title not available (Why is that?)
• Non-Parametric Estimation II. Statistically Equivalent Blocks and Tolerance Regions--The Continuous Case
• Nonparametric Estimation, III. Statistically Equivalent Blocks and Multivariate Tolerance Regions--The Discontinuous Case
• Nonparametric Estimation IV
• Non-Parametric Estimation. I. Validation of Order Statistics

Cited In (17)

• Criteria of efficiency for set-valued classification
• A fast conformal classifier based on multi-output extreme learning machine
• Distribution-free predictive inference for regression
• Conformal prediction under hypergraphical models
• Conformal prediction: a unified review of theory and new challenges
• Cross-conformal predictors
• Inductive Venn prediction
• Unsupervised streaming anomaly detection for instrumented infrastructure
• Classification with minimum ambiguity under distribution heterogeneity
• Efficient and multiply robust risk estimation under general forms of dataset shift
• Least Ambiguous Set-Valued Classifiers With Bounded Error Levels
• Training-conditional coverage for distribution-free predictive inference
• Title not available (Why is that?)
• Valid model-free prediction of future insurance claims
• Majority vote ensembles of conformal predictors
• Combination of inductive Mondrian conformal predictors
• Conditional predictive inference for stable algorithms

Uses Software

• Digital Library of Mathematical Functions
• UCI-ml
• DAAG
• AdaBoost.MH
• ElemStatLearn
• DAAGxtras

This page was built for publication: Conditional validity of inductive conformal predictors

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q374136)