Estimation of the order of an auto-regressive model

Sudarshan Rao Nelatury, P. S. Moharir

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

There are no definitive solutions available for the order estimation of the auto-regressive process. First, the performance of the three criteria, namely FPE, AIC and MDL is illustrated. Next, it is indicated that there are possibilities of their performance being improved. The algorithm tri proposed here utilizes three minimum values of any of the conventional loss functions in the FPE, AIC and MDL methods. It also uses three statistics derived from these three minimum values. The estimated order is a rounded weighted average of these six statistics. The algorithm is found to do better in a qualified sense of yielding peakier distribution of the estimated orders when tested for 1000 synthetic models of orders 3, 5 and 7 each. The conclusion drawn is that there are open possibilities of improving upon the conventional order estimators for auto-regressive processes. This means that till axiologically sounder estimators are available one should be ready to use heuristic estimators proposed here.

Original languageEnglish (US)
Pages (from-to)749-758
Number of pages10
JournalSadhana
Volume20
Issue number5
DOIs
StatePublished - Oct 1 1995

Fingerprint

Statistics
Acoustic waves

All Science Journal Classification (ASJC) codes

  • General

Cite this

Nelatury, Sudarshan Rao ; Moharir, P. S. / Estimation of the order of an auto-regressive model. In: Sadhana. 1995 ; Vol. 20, No. 5. pp. 749-758.
@article{b78420d70755497397182204412fff75,
title = "Estimation of the order of an auto-regressive model",
abstract = "There are no definitive solutions available for the order estimation of the auto-regressive process. First, the performance of the three criteria, namely FPE, AIC and MDL is illustrated. Next, it is indicated that there are possibilities of their performance being improved. The algorithm tri proposed here utilizes three minimum values of any of the conventional loss functions in the FPE, AIC and MDL methods. It also uses three statistics derived from these three minimum values. The estimated order is a rounded weighted average of these six statistics. The algorithm is found to do better in a qualified sense of yielding peakier distribution of the estimated orders when tested for 1000 synthetic models of orders 3, 5 and 7 each. The conclusion drawn is that there are open possibilities of improving upon the conventional order estimators for auto-regressive processes. This means that till axiologically sounder estimators are available one should be ready to use heuristic estimators proposed here.",
author = "Nelatury, {Sudarshan Rao} and Moharir, {P. S.}",
year = "1995",
month = "10",
day = "1",
doi = "10.1007/BF02744408",
language = "English (US)",
volume = "20",
pages = "749--758",
journal = "Sadhana - Academy Proceedings in Engineering Sciences",
issn = "0256-2499",
publisher = "Springer India",
number = "5",

}

Estimation of the order of an auto-regressive model. / Nelatury, Sudarshan Rao; Moharir, P. S.

In: Sadhana, Vol. 20, No. 5, 01.10.1995, p. 749-758.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Estimation of the order of an auto-regressive model

AU - Nelatury, Sudarshan Rao

AU - Moharir, P. S.

PY - 1995/10/1

Y1 - 1995/10/1

N2 - There are no definitive solutions available for the order estimation of the auto-regressive process. First, the performance of the three criteria, namely FPE, AIC and MDL is illustrated. Next, it is indicated that there are possibilities of their performance being improved. The algorithm tri proposed here utilizes three minimum values of any of the conventional loss functions in the FPE, AIC and MDL methods. It also uses three statistics derived from these three minimum values. The estimated order is a rounded weighted average of these six statistics. The algorithm is found to do better in a qualified sense of yielding peakier distribution of the estimated orders when tested for 1000 synthetic models of orders 3, 5 and 7 each. The conclusion drawn is that there are open possibilities of improving upon the conventional order estimators for auto-regressive processes. This means that till axiologically sounder estimators are available one should be ready to use heuristic estimators proposed here.

AB - There are no definitive solutions available for the order estimation of the auto-regressive process. First, the performance of the three criteria, namely FPE, AIC and MDL is illustrated. Next, it is indicated that there are possibilities of their performance being improved. The algorithm tri proposed here utilizes three minimum values of any of the conventional loss functions in the FPE, AIC and MDL methods. It also uses three statistics derived from these three minimum values. The estimated order is a rounded weighted average of these six statistics. The algorithm is found to do better in a qualified sense of yielding peakier distribution of the estimated orders when tested for 1000 synthetic models of orders 3, 5 and 7 each. The conclusion drawn is that there are open possibilities of improving upon the conventional order estimators for auto-regressive processes. This means that till axiologically sounder estimators are available one should be ready to use heuristic estimators proposed here.

UR - http://www.scopus.com/inward/record.url?scp=51249164848&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=51249164848&partnerID=8YFLogxK

U2 - 10.1007/BF02744408

DO - 10.1007/BF02744408

M3 - Article

AN - SCOPUS:51249164848

VL - 20

SP - 749

EP - 758

JO - Sadhana - Academy Proceedings in Engineering Sciences

JF - Sadhana - Academy Proceedings in Engineering Sciences

SN - 0256-2499

IS - 5

ER -