Product Entropy of Gaussian Distributions

This paper studies the product epsilon entropy of mean-continuous Gaussian processes. That is, a given mean-continuous Gaussian process on the unit interval is expanded into its Karhunen expansion. Along the kth eigenfunction axis, a partition by intervals of length εkis made, and the entropy of the...

Full description

Saved in:
Bibliographic Details
Published in:The Annals of mathematical statistics 1969-06, Vol.40 (3), p.870-904
Main Authors: Posner, Edward C., Rodemich, Eugene R., Rumsey, Howard
Format: Article
Language:English
Subjects:
Eigenvalues
Entropy
Mathematical functions
Mathematical intervals
Mathematics
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page 904
container_issue 3
container_start_page 870
container_title The Annals of mathematical statistics
container_volume 40
creator Posner, Edward C.
Rodemich, Eugene R.
Rumsey, Howard
description This paper studies the product epsilon entropy of mean-continuous Gaussian processes. That is, a given mean-continuous Gaussian process on the unit interval is expanded into its Karhunen expansion. Along the kth eigenfunction axis, a partition by intervals of length εkis made, and the entropy of the resulting discrete distribution is noted. The infimum of the sum over k of these entropies subject to the constraint that ∑ εk 2≤ ε2is the product epsilon entropy of the process. It is shown that the best partition to take along each eigenfunction axis is the one in which 0 is the midpoint of an interval in the partition. Furthermore, the product epsilon entropy is finite if and only if ∑ λklog λk -1is finite, where λkis the kth eigenvalue of the process. When the above series is finite, the values of εkwhich achieve the product entropy are found. Asymptotic expressions for the product epsilon entropy are derived in some special cases. The problem arises in the theory of data compression, which studies the efficient representation of random data with prescribed accuracy
doi_str_mv 10.1214/aoms/1177697595
format article
fullrecord <record><control><sourceid>jstor_proje</sourceid><recordid>TN_cdi_projecteuclid_primary_oai_CULeuclid_euclid_aoms_1177697595</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><jstor_id>2239636</jstor_id><sourcerecordid>2239636</sourcerecordid><originalsourceid>FETCH-LOGICAL-c245t-28bb283307a2df9f1065d85e9cc3351185a0622c8cab290b464fbc02c82278733</originalsourceid><addsrcrecordid>eNptkDFrwzAUhEVpoW7auUsHrx1cS0-WLI3BTdKCoR2a2ciyDTKJFSR5yL9vjE26dDre8b6DO4SeCX4jQLJU2aNPCclzLnMm2Q2KgHCRCCnxLYowxjTJBCP36MH7fjqp4BF6_Xa2GXWIN0Nw9nSObRfv1Oi9UUP8bnxwph6DsYN_RHedOvj2adEV2m83P8VHUn7tPot1mWjIWEhA1DUISnGuoOlkRzBnjWCt1JpSRohgCnMALbSqQeI641lXa3wxAHKRU7pC6zn35Gzf6tCO-mCa6uTMUblzZZWpin25uItM3au_7peMdM7Qznrv2u6KE1xNa_1DvMxE74N113cAKjnl9Bdt-2ca</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Product Entropy of Gaussian Distributions</title><source>JSTOR Archival Journals and Primary Sources Collection</source><source>Project Euclid Open Access Journals</source><creator>Posner, Edward C. ; Rodemich, Eugene R. ; Rumsey, Howard</creator><creatorcontrib>Posner, Edward C. ; Rodemich, Eugene R. ; Rumsey, Howard</creatorcontrib><description>This paper studies the product epsilon entropy of mean-continuous Gaussian processes. That is, a given mean-continuous Gaussian process on the unit interval is expanded into its Karhunen expansion. Along the kth eigenfunction axis, a partition by intervals of length εkis made, and the entropy of the resulting discrete distribution is noted. The infimum of the sum over k of these entropies subject to the constraint that ∑ εk 2≤ ε2is the product epsilon entropy of the process. It is shown that the best partition to take along each eigenfunction axis is the one in which 0 is the midpoint of an interval in the partition. Furthermore, the product epsilon entropy is finite if and only if ∑ λklog λk -1is finite, where λkis the kth eigenvalue of the process. When the above series is finite, the values of εkwhich achieve the product entropy are found. Asymptotic expressions for the product epsilon entropy are derived in some special cases. The problem arises in the theory of data compression, which studies the efficient representation of random data with prescribed accuracy</description><identifier>ISSN: 0003-4851</identifier><identifier>EISSN: 2168-8990</identifier><identifier>DOI: 10.1214/aoms/1177697595</identifier><language>eng</language><publisher>Institute of Mathematical Statistics</publisher><subject>Eigenvalues ; Entropy ; Mathematical functions ; Mathematical intervals ; Mathematics</subject><ispartof>The Annals of mathematical statistics, 1969-06, Vol.40 (3), p.870-904</ispartof><rights>Copyright 1969 Institute of Mathematical Statistics</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://www.jstor.org/stable/pdf/2239636$$EPDF$$P50$$Gjstor$$H</linktopdf><linktohtml>$$Uhttps://www.jstor.org/stable/2239636$$EHTML$$P50$$Gjstor$$H</linktohtml><link.rule.ids>230,314,776,780,878,881,27901,27902,58213,58446,79726,79727</link.rule.ids></links><search><creatorcontrib>Posner, Edward C.</creatorcontrib><creatorcontrib>Rodemich, Eugene R.</creatorcontrib><creatorcontrib>Rumsey, Howard</creatorcontrib><title>Product Entropy of Gaussian Distributions</title><title>The Annals of mathematical statistics</title><description>This paper studies the product epsilon entropy of mean-continuous Gaussian processes. That is, a given mean-continuous Gaussian process on the unit interval is expanded into its Karhunen expansion. Along the kth eigenfunction axis, a partition by intervals of length εkis made, and the entropy of the resulting discrete distribution is noted. The infimum of the sum over k of these entropies subject to the constraint that ∑ εk 2≤ ε2is the product epsilon entropy of the process. It is shown that the best partition to take along each eigenfunction axis is the one in which 0 is the midpoint of an interval in the partition. Furthermore, the product epsilon entropy is finite if and only if ∑ λklog λk -1is finite, where λkis the kth eigenvalue of the process. When the above series is finite, the values of εkwhich achieve the product entropy are found. Asymptotic expressions for the product epsilon entropy are derived in some special cases. The problem arises in the theory of data compression, which studies the efficient representation of random data with prescribed accuracy</description><subject>Eigenvalues</subject><subject>Entropy</subject><subject>Mathematical functions</subject><subject>Mathematical intervals</subject><subject>Mathematics</subject><issn>0003-4851</issn><issn>2168-8990</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>1969</creationdate><recordtype>article</recordtype><recordid>eNptkDFrwzAUhEVpoW7auUsHrx1cS0-WLI3BTdKCoR2a2ciyDTKJFSR5yL9vjE26dDre8b6DO4SeCX4jQLJU2aNPCclzLnMm2Q2KgHCRCCnxLYowxjTJBCP36MH7fjqp4BF6_Xa2GXWIN0Nw9nSObRfv1Oi9UUP8bnxwph6DsYN_RHedOvj2adEV2m83P8VHUn7tPot1mWjIWEhA1DUISnGuoOlkRzBnjWCt1JpSRohgCnMALbSqQeI641lXa3wxAHKRU7pC6zn35Gzf6tCO-mCa6uTMUblzZZWpin25uItM3au_7peMdM7Qznrv2u6KE1xNa_1DvMxE74N113cAKjnl9Bdt-2ca</recordid><startdate>19690601</startdate><enddate>19690601</enddate><creator>Posner, Edward C.</creator><creator>Rodemich, Eugene R.</creator><creator>Rumsey, Howard</creator><general>Institute of Mathematical Statistics</general><general>The Institute of Mathematical Statistics</general><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>19690601</creationdate><title>Product Entropy of Gaussian Distributions</title><author>Posner, Edward C. ; Rodemich, Eugene R. ; Rumsey, Howard</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c245t-28bb283307a2df9f1065d85e9cc3351185a0622c8cab290b464fbc02c82278733</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>1969</creationdate><topic>Eigenvalues</topic><topic>Entropy</topic><topic>Mathematical functions</topic><topic>Mathematical intervals</topic><topic>Mathematics</topic><toplevel>online_resources</toplevel><creatorcontrib>Posner, Edward C.</creatorcontrib><creatorcontrib>Rodemich, Eugene R.</creatorcontrib><creatorcontrib>Rumsey, Howard</creatorcontrib><collection>CrossRef</collection><jtitle>The Annals of mathematical statistics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Posner, Edward C.</au><au>Rodemich, Eugene R.</au><au>Rumsey, Howard</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Product Entropy of Gaussian Distributions</atitle><jtitle>The Annals of mathematical statistics</jtitle><date>1969-06-01</date><risdate>1969</risdate><volume>40</volume><issue>3</issue><spage>870</spage><epage>904</epage><pages>870-904</pages><issn>0003-4851</issn><eissn>2168-8990</eissn><abstract>This paper studies the product epsilon entropy of mean-continuous Gaussian processes. That is, a given mean-continuous Gaussian process on the unit interval is expanded into its Karhunen expansion. Along the kth eigenfunction axis, a partition by intervals of length εkis made, and the entropy of the resulting discrete distribution is noted. The infimum of the sum over k of these entropies subject to the constraint that ∑ εk 2≤ ε2is the product epsilon entropy of the process. It is shown that the best partition to take along each eigenfunction axis is the one in which 0 is the midpoint of an interval in the partition. Furthermore, the product epsilon entropy is finite if and only if ∑ λklog λk -1is finite, where λkis the kth eigenvalue of the process. When the above series is finite, the values of εkwhich achieve the product entropy are found. Asymptotic expressions for the product epsilon entropy are derived in some special cases. The problem arises in the theory of data compression, which studies the efficient representation of random data with prescribed accuracy</abstract><pub>Institute of Mathematical Statistics</pub><doi>10.1214/aoms/1177697595</doi><tpages>35</tpages><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier ISSN: 0003-4851
ispartof The Annals of mathematical statistics, 1969-06, Vol.40 (3), p.870-904
issn 0003-4851
2168-8990
language eng
recordid cdi_projecteuclid_primary_oai_CULeuclid_euclid_aoms_1177697595
source JSTOR Archival Journals and Primary Sources Collection; Project Euclid Open Access Journals
subjects Eigenvalues
Entropy
Mathematical functions
Mathematical intervals
Mathematics
title Product Entropy of Gaussian Distributions
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-24T00%3A05%3A44IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-jstor_proje&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Product%20Entropy%20of%20Gaussian%20Distributions&rft.jtitle=The%20Annals%20of%20mathematical%20statistics&rft.au=Posner,%20Edward%20C.&rft.date=1969-06-01&rft.volume=40&rft.issue=3&rft.spage=870&rft.epage=904&rft.pages=870-904&rft.issn=0003-4851&rft.eissn=2168-8990&rft_id=info:doi/10.1214/aoms/1177697595&rft_dat=%3Cjstor_proje%3E2239636%3C/jstor_proje%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c245t-28bb283307a2df9f1065d85e9cc3351185a0622c8cab290b464fbc02c82278733%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_jstor_id=2239636&rfr_iscdi=true