One-unit contrast functions for independent component analysis: a statistical analysis

The author (1997) introduced a large family of one-unit contrast functions to be used in independent component analysis (ICA). In this paper, the family is analyzed mathematically in the case of a finite sample. Two aspects of the estimators obtained using such contrast functions are considered: asy...

Full description

Saved in:
Bibliographic Details
Main Author: Hyvarinen, A.
Format: Conference Proceeding
Language:English
Subjects:
Blind source separation
Covariance matrix
Gaussian noise
Independent component analysis
Information science
Probability
Robustness
Signal processing
Statistical analysis
Vectors
Citations: Items that cite this one
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by cdi_FETCH-LOGICAL-c151t-ed5acee2fd29fe582399f8ad945068171428e81500543a07d72bba0e282edf653
cites
container_end_page 397
container_issue
container_start_page 388
container_title
container_volume
creator Hyvarinen, A.
description The author (1997) introduced a large family of one-unit contrast functions to be used in independent component analysis (ICA). In this paper, the family is analyzed mathematically in the case of a finite sample. Two aspects of the estimators obtained using such contrast functions are considered: asymptotic variance, and robustness against outliers. An expression for the contrast function that minimizes the asymptotic variance is obtained as a function of the probability densities of the independent components. Combined with robustness considerations, these results provide strong arguments in favor of the use of contrast functions based on slowly growing functions, and against the use of kurtosis, which is the classical contrast function.
doi_str_mv 10.1109/NNSP.1997.622420
format conference_proceeding
fullrecord <record><control><sourceid>ieee_6IE</sourceid><recordid>TN_cdi_ieee_primary_622420</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>622420</ieee_id><sourcerecordid>622420</sourcerecordid><originalsourceid>FETCH-LOGICAL-c151t-ed5acee2fd29fe582399f8ad945068171428e81500543a07d72bba0e282edf653</originalsourceid><addsrcrecordid>eNo9kE1LAzEYhIMfYFu9i6f8gV3fvNlsEm9SrAqlFfy4lnT3DUS22WWTHvrvtVS8zAwPwxyGsVsBpRBg71er97dSWKvLGrFCOGMTlNoWKNGesyloA7JCVdsLNhFgbCGVUldsmtI3AALqesK-1pGKfQyZN33Mo0uZ-31scuhj4r4feYgtDfQr8VjZDX08Jhddd0ghPXDHU3Y5pBwa1_3za3bpXZfo5s9n7HPx9DF_KZbr59f547JohBK5oFa5hgh9i9aTMiit9ca1tlJQG6FFhYaMUACqkg50q3G7dUBokFpfKzljd6fdQESbYQw7Nx42pzvkD3NwU2I</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>One-unit contrast functions for independent component analysis: a statistical analysis</title><source>IEEE Electronic Library (IEL) Conference Proceedings</source><creator>Hyvarinen, A.</creator><creatorcontrib>Hyvarinen, A.</creatorcontrib><description>The author (1997) introduced a large family of one-unit contrast functions to be used in independent component analysis (ICA). In this paper, the family is analyzed mathematically in the case of a finite sample. Two aspects of the estimators obtained using such contrast functions are considered: asymptotic variance, and robustness against outliers. An expression for the contrast function that minimizes the asymptotic variance is obtained as a function of the probability densities of the independent components. Combined with robustness considerations, these results provide strong arguments in favor of the use of contrast functions based on slowly growing functions, and against the use of kurtosis, which is the classical contrast function.</description><identifier>ISSN: 1089-3555</identifier><identifier>ISBN: 0780342569</identifier><identifier>ISBN: 9780780342569</identifier><identifier>EISSN: 2379-2329</identifier><identifier>DOI: 10.1109/NNSP.1997.622420</identifier><language>eng</language><publisher>IEEE</publisher><subject>Blind source separation ; Covariance matrix ; Gaussian noise ; Independent component analysis ; Information science ; Probability ; Robustness ; Signal processing ; Statistical analysis ; Vectors</subject><ispartof>Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop, 1997, p.388-397</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c151t-ed5acee2fd29fe582399f8ad945068171428e81500543a07d72bba0e282edf653</citedby></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/622420$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,777,781,786,787,2052,4036,4037,27906,54536,54901,54913</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/622420$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Hyvarinen, A.</creatorcontrib><title>One-unit contrast functions for independent component analysis: a statistical analysis</title><title>Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop</title><addtitle>NNSP</addtitle><description>The author (1997) introduced a large family of one-unit contrast functions to be used in independent component analysis (ICA). In this paper, the family is analyzed mathematically in the case of a finite sample. Two aspects of the estimators obtained using such contrast functions are considered: asymptotic variance, and robustness against outliers. An expression for the contrast function that minimizes the asymptotic variance is obtained as a function of the probability densities of the independent components. Combined with robustness considerations, these results provide strong arguments in favor of the use of contrast functions based on slowly growing functions, and against the use of kurtosis, which is the classical contrast function.</description><subject>Blind source separation</subject><subject>Covariance matrix</subject><subject>Gaussian noise</subject><subject>Independent component analysis</subject><subject>Information science</subject><subject>Probability</subject><subject>Robustness</subject><subject>Signal processing</subject><subject>Statistical analysis</subject><subject>Vectors</subject><issn>1089-3555</issn><issn>2379-2329</issn><isbn>0780342569</isbn><isbn>9780780342569</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>1997</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNo9kE1LAzEYhIMfYFu9i6f8gV3fvNlsEm9SrAqlFfy4lnT3DUS22WWTHvrvtVS8zAwPwxyGsVsBpRBg71er97dSWKvLGrFCOGMTlNoWKNGesyloA7JCVdsLNhFgbCGVUldsmtI3AALqesK-1pGKfQyZN33Mo0uZ-31scuhj4r4feYgtDfQr8VjZDX08Jhddd0ghPXDHU3Y5pBwa1_3za3bpXZfo5s9n7HPx9DF_KZbr59f547JohBK5oFa5hgh9i9aTMiit9ca1tlJQG6FFhYaMUACqkg50q3G7dUBokFpfKzljd6fdQESbYQw7Nx42pzvkD3NwU2I</recordid><startdate>1997</startdate><enddate>1997</enddate><creator>Hyvarinen, A.</creator><general>IEEE</general><scope>6IE</scope><scope>6IL</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIL</scope></search><sort><creationdate>1997</creationdate><title>One-unit contrast functions for independent component analysis: a statistical analysis</title><author>Hyvarinen, A.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c151t-ed5acee2fd29fe582399f8ad945068171428e81500543a07d72bba0e282edf653</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>1997</creationdate><topic>Blind source separation</topic><topic>Covariance matrix</topic><topic>Gaussian noise</topic><topic>Independent component analysis</topic><topic>Information science</topic><topic>Probability</topic><topic>Robustness</topic><topic>Signal processing</topic><topic>Statistical analysis</topic><topic>Vectors</topic><toplevel>online_resources</toplevel><creatorcontrib>Hyvarinen, A.</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Xplore</collection><collection>IEEE Proceedings Order Plans (POP All) 1998-Present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Hyvarinen, A.</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>One-unit contrast functions for independent component analysis: a statistical analysis</atitle><btitle>Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop</btitle><stitle>NNSP</stitle><date>1997</date><risdate>1997</risdate><spage>388</spage><epage>397</epage><pages>388-397</pages><issn>1089-3555</issn><eissn>2379-2329</eissn><isbn>0780342569</isbn><isbn>9780780342569</isbn><abstract>The author (1997) introduced a large family of one-unit contrast functions to be used in independent component analysis (ICA). In this paper, the family is analyzed mathematically in the case of a finite sample. Two aspects of the estimators obtained using such contrast functions are considered: asymptotic variance, and robustness against outliers. An expression for the contrast function that minimizes the asymptotic variance is obtained as a function of the probability densities of the independent components. Combined with robustness considerations, these results provide strong arguments in favor of the use of contrast functions based on slowly growing functions, and against the use of kurtosis, which is the classical contrast function.</abstract><pub>IEEE</pub><doi>10.1109/NNSP.1997.622420</doi><tpages>10</tpages></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 1089-3555
ispartof Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop, 1997, p.388-397
issn 1089-3555
2379-2329
language eng
recordid cdi_ieee_primary_622420
source IEEE Electronic Library (IEL) Conference Proceedings
subjects Blind source separation
Covariance matrix
Gaussian noise
Independent component analysis
Information science
Probability
Robustness
Signal processing
Statistical analysis
Vectors
title One-unit contrast functions for independent component analysis: a statistical analysis
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-20T05%3A29%3A28IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_6IE&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=One-unit%20contrast%20functions%20for%20independent%20component%20analysis:%20a%20statistical%20analysis&rft.btitle=Neural%20Networks%20for%20Signal%20Processing%20VII.%20Proceedings%20of%20the%201997%20IEEE%20Signal%20Processing%20Society%20Workshop&rft.au=Hyvarinen,%20A.&rft.date=1997&rft.spage=388&rft.epage=397&rft.pages=388-397&rft.issn=1089-3555&rft.eissn=2379-2329&rft.isbn=0780342569&rft.isbn_list=9780780342569&rft_id=info:doi/10.1109/NNSP.1997.622420&rft_dat=%3Cieee_6IE%3E622420%3C/ieee_6IE%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c151t-ed5acee2fd29fe582399f8ad945068171428e81500543a07d72bba0e282edf653%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=622420&rfr_iscdi=true