Limitations on intermittent forecasting

G. Morvai, Benjamin Weiss

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

Bailey showed that the general pointwise forecasting for stationary and ergodic time series has a negative solution. However, it is known that for Markov chains the problem can be solved. Morvai showed that there is a stopping time sequence {λn} such that P(Xλn+1 = 1 X0,..., Xλn) can be estimated from samples (X0,..., Xλn) such that the difference between the conditional probability and the estimate vanishes along these stoppping times for all stationary and ergodic binary time series. We will show it is not possible to estimate the above conditional probability along a stopping time sequence for all stationary and ergodic binary time series in a pointwise sense such that if the time series turns out to be a Markov chain, the predictor will predict eventually for all n.

Original languageEnglish
Pages (from-to)285-290
Number of pages6
JournalStatistics and Probability Letters
Volume72
Issue number4
DOIs
Publication statusPublished - May 15 2005

Fingerprint

Forecasting
Time series
Stopping Time
Conditional probability
Markov chain
Binary
Estimate
Predictors
Vanish
Predict
Stopping time

Keywords

  • Finite-order Markov chains
  • Nonparametric estimation
  • Prediction theory
  • Stationary and ergodic processes

ASJC Scopus subject areas

  • Statistics, Probability and Uncertainty
  • Statistics and Probability

Cite this

Limitations on intermittent forecasting. / Morvai, G.; Weiss, Benjamin.

In: Statistics and Probability Letters, Vol. 72, No. 4, 15.05.2005, p. 285-290.

Research output: Contribution to journalArticle

Morvai, G. ; Weiss, Benjamin. / Limitations on intermittent forecasting. In: Statistics and Probability Letters. 2005 ; Vol. 72, No. 4. pp. 285-290.
@article{6b40571f85fe4a1e8a4e11022476142f,
title = "Limitations on intermittent forecasting",
abstract = "Bailey showed that the general pointwise forecasting for stationary and ergodic time series has a negative solution. However, it is known that for Markov chains the problem can be solved. Morvai showed that there is a stopping time sequence {λn} such that P(Xλn+1 = 1 X0,..., Xλn) can be estimated from samples (X0,..., Xλn) such that the difference between the conditional probability and the estimate vanishes along these stoppping times for all stationary and ergodic binary time series. We will show it is not possible to estimate the above conditional probability along a stopping time sequence for all stationary and ergodic binary time series in a pointwise sense such that if the time series turns out to be a Markov chain, the predictor will predict eventually for all n.",
keywords = "Finite-order Markov chains, Nonparametric estimation, Prediction theory, Stationary and ergodic processes",
author = "G. Morvai and Benjamin Weiss",
year = "2005",
month = "5",
day = "15",
doi = "10.1016/j.spl.2004.12.016",
language = "English",
volume = "72",
pages = "285--290",
journal = "Statistics and Probability Letters",
issn = "0167-7152",
publisher = "Elsevier",
number = "4",

}

TY - JOUR

T1 - Limitations on intermittent forecasting

AU - Morvai, G.

AU - Weiss, Benjamin

PY - 2005/5/15

Y1 - 2005/5/15

N2 - Bailey showed that the general pointwise forecasting for stationary and ergodic time series has a negative solution. However, it is known that for Markov chains the problem can be solved. Morvai showed that there is a stopping time sequence {λn} such that P(Xλn+1 = 1 X0,..., Xλn) can be estimated from samples (X0,..., Xλn) such that the difference between the conditional probability and the estimate vanishes along these stoppping times for all stationary and ergodic binary time series. We will show it is not possible to estimate the above conditional probability along a stopping time sequence for all stationary and ergodic binary time series in a pointwise sense such that if the time series turns out to be a Markov chain, the predictor will predict eventually for all n.

AB - Bailey showed that the general pointwise forecasting for stationary and ergodic time series has a negative solution. However, it is known that for Markov chains the problem can be solved. Morvai showed that there is a stopping time sequence {λn} such that P(Xλn+1 = 1 X0,..., Xλn) can be estimated from samples (X0,..., Xλn) such that the difference between the conditional probability and the estimate vanishes along these stoppping times for all stationary and ergodic binary time series. We will show it is not possible to estimate the above conditional probability along a stopping time sequence for all stationary and ergodic binary time series in a pointwise sense such that if the time series turns out to be a Markov chain, the predictor will predict eventually for all n.

KW - Finite-order Markov chains

KW - Nonparametric estimation

KW - Prediction theory

KW - Stationary and ergodic processes

UR - http://www.scopus.com/inward/record.url?scp=17444383424&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=17444383424&partnerID=8YFLogxK

U2 - 10.1016/j.spl.2004.12.016

DO - 10.1016/j.spl.2004.12.016

M3 - Article

AN - SCOPUS:17444383424

VL - 72

SP - 285

EP - 290

JO - Statistics and Probability Letters

JF - Statistics and Probability Letters

SN - 0167-7152

IS - 4

ER -