A Justification of Conditional Confidence Intervals

Research output: Working paperProfessional

77 Downloads (Pure)

Abstract

To quantify uncertainty around point estimates of conditional objects such as conditional means or variances, parameter uncertainty has to be taken into account. Attempts to incorporate parameter uncertainty are typically based on the unrealistic assumption of observing two independent processes, where one is used for parameter estimation, and the other for conditioning upon. Such unrealistic foundation raises the question whether these intervals are theoretically justified in a realistic setting. This paper presents an asymptotic justification for this type of intervals that does not require such an unrealistic assumption, but relies on a sample-split approach instead. By showing that our sample-split intervals coincide asymptotically with the standard intervals, we provide a novel, and realistic, justification for confidence intervals of conditional objects. The analysis is carried out for a general class of Markov chains nesting various time series models.
Original languageEnglish
Number of pages54
Publication statusPublished - 2 Oct 2017

Keywords

  • Conditional confidence intervals
  • Parameter uncertainty
  • Markov chains
  • Prediction
  • Sample-splitting
  • Merging

Cite this

@techreport{8735997f655b4840953d1ad5a905d642,
title = "A Justification of Conditional Confidence Intervals",
abstract = "To quantify uncertainty around point estimates of conditional objects such as conditional means or variances, parameter uncertainty has to be taken into account. Attempts to incorporate parameter uncertainty are typically based on the unrealistic assumption of observing two independent processes, where one is used for parameter estimation, and the other for conditioning upon. Such unrealistic foundation raises the question whether these intervals are theoretically justified in a realistic setting. This paper presents an asymptotic justification for this type of intervals that does not require such an unrealistic assumption, but relies on a sample-split approach instead. By showing that our sample-split intervals coincide asymptotically with the standard intervals, we provide a novel, and realistic, justification for confidence intervals of conditional objects. The analysis is carried out for a general class of Markov chains nesting various time series models.",
keywords = "Conditional confidence intervals, Parameter uncertainty, Markov chains, Prediction, Sample-splitting, Merging",
author = "Eric Beutner and Alexander Heinemann and Stephan Smeekes",
year = "2017",
month = "10",
day = "2",
language = "English",
type = "WorkingPaper",

}

A Justification of Conditional Confidence Intervals. / Beutner, Eric; Heinemann, Alexander; Smeekes, Stephan.

2017.

Research output: Working paperProfessional

TY - UNPB

T1 - A Justification of Conditional Confidence Intervals

AU - Beutner, Eric

AU - Heinemann, Alexander

AU - Smeekes, Stephan

PY - 2017/10/2

Y1 - 2017/10/2

N2 - To quantify uncertainty around point estimates of conditional objects such as conditional means or variances, parameter uncertainty has to be taken into account. Attempts to incorporate parameter uncertainty are typically based on the unrealistic assumption of observing two independent processes, where one is used for parameter estimation, and the other for conditioning upon. Such unrealistic foundation raises the question whether these intervals are theoretically justified in a realistic setting. This paper presents an asymptotic justification for this type of intervals that does not require such an unrealistic assumption, but relies on a sample-split approach instead. By showing that our sample-split intervals coincide asymptotically with the standard intervals, we provide a novel, and realistic, justification for confidence intervals of conditional objects. The analysis is carried out for a general class of Markov chains nesting various time series models.

AB - To quantify uncertainty around point estimates of conditional objects such as conditional means or variances, parameter uncertainty has to be taken into account. Attempts to incorporate parameter uncertainty are typically based on the unrealistic assumption of observing two independent processes, where one is used for parameter estimation, and the other for conditioning upon. Such unrealistic foundation raises the question whether these intervals are theoretically justified in a realistic setting. This paper presents an asymptotic justification for this type of intervals that does not require such an unrealistic assumption, but relies on a sample-split approach instead. By showing that our sample-split intervals coincide asymptotically with the standard intervals, we provide a novel, and realistic, justification for confidence intervals of conditional objects. The analysis is carried out for a general class of Markov chains nesting various time series models.

KW - Conditional confidence intervals

KW - Parameter uncertainty

KW - Markov chains

KW - Prediction

KW - Sample-splitting

KW - Merging

M3 - Working paper

BT - A Justification of Conditional Confidence Intervals

ER -