Application-Dedicated Selection of Filters (ADSF) using covariance maximization and orthogonal projection.
Identifieur interne : 001F06 ( PubMed/Curation ); précédent : 001F05; suivant : 001F07Application-Dedicated Selection of Filters (ADSF) using covariance maximization and orthogonal projection.
Auteurs : Xavier Hadoux [Australie] ; Dinesh Kant Kumar [Australie] ; Marc G. Sarossy [Australie] ; Jean-Michel Roger [France] ; Nathalie Gorretta [France]Source :
- Analytica chimica acta [ 1873-4324 ] ; 2016.
Abstract
Visible and near-infrared (Vis-NIR) spectra are generated by the combination of numerous low resolution features. Spectral variables are thus highly correlated, which can cause problems for selecting the most appropriate ones for a given application. Some decomposition bases such as Fourier or wavelet generally help highlighting spectral features that are important, but are by nature constraint to have both positive and negative components. Thus, in addition to complicating the selected features interpretability, it impedes their use for application-dedicated sensors. In this paper we have proposed a new method for feature selection: Application-Dedicated Selection of Filters (ADSF). This method relaxes the shape constraint by enabling the selection of any type of user defined custom features. By considering only relevant features, based on the underlying nature of the data, high regularization of the final model can be obtained, even in the small sample size context often encountered in spectroscopic applications. For larger scale deployment of application-dedicated sensors, these predefined feature constraints can lead to application specific optical filters, e.g., lowpass, highpass, bandpass or bandstop filters with positive only coefficients. In a similar fashion to Partial Least Squares, ADSF successively selects features using covariance maximization and deflates their influences using orthogonal projection in order to optimally tune the selection to the data with limited redundancy. ADSF is well suited for spectroscopic data as it can deal with large numbers of highly correlated variables in supervised learning, even with many correlated responses.
DOI: 10.1016/j.aca.2016.04.004
PubMed: 27126785
Links toward previous steps (curation, corpus...)
- to stream PubMed, to step Corpus: Pour aller vers cette notice dans l'étape Curation :001F30
Links to Exploration step
pubmed:27126785Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">Application-Dedicated Selection of Filters (ADSF) using covariance maximization and orthogonal projection.</title><author><name sortKey="Hadoux, Xavier" sort="Hadoux, Xavier" uniqKey="Hadoux X" first="Xavier" last="Hadoux">Xavier Hadoux</name><affiliation wicri:level="1"><nlm:affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France; Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia. Electronic address: xavier@hadoux.com.</nlm:affiliation><country xml:lang="fr">Australie</country><wicri:regionArea>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France; Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery</wicri:regionArea></affiliation></author><author><name sortKey="Kumar, Dinesh Kant" sort="Kumar, Dinesh Kant" uniqKey="Kumar D" first="Dinesh Kant" last="Kumar">Dinesh Kant Kumar</name><affiliation wicri:level="1"><nlm:affiliation>Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001, Australia.</nlm:affiliation><country xml:lang="fr">Australie</country><wicri:regionArea>Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001</wicri:regionArea></affiliation></author><author><name sortKey="Sarossy, Marc G" sort="Sarossy, Marc G" uniqKey="Sarossy M" first="Marc G" last="Sarossy">Marc G. Sarossy</name><affiliation wicri:level="1"><nlm:affiliation>Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia; Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001, Australia.</nlm:affiliation><country xml:lang="fr">Australie</country><wicri:regionArea>Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia; Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001</wicri:regionArea></affiliation></author><author><name sortKey="Roger, Jean Michel" sort="Roger, Jean Michel" uniqKey="Roger J" first="Jean-Michel" last="Roger">Jean-Michel Roger</name><affiliation wicri:level="1"><nlm:affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France.</nlm:affiliation><country xml:lang="fr">France</country><wicri:regionArea>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5</wicri:regionArea></affiliation></author><author><name sortKey="Gorretta, Nathalie" sort="Gorretta, Nathalie" uniqKey="Gorretta N" first="Nathalie" last="Gorretta">Nathalie Gorretta</name><affiliation wicri:level="1"><nlm:affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France.</nlm:affiliation><country xml:lang="fr">France</country><wicri:regionArea>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5</wicri:regionArea></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">PubMed</idno><date when="2016">2016</date><idno type="RBID">pubmed:27126785</idno><idno type="pmid">27126785</idno><idno type="doi">10.1016/j.aca.2016.04.004</idno><idno type="wicri:Area/PubMed/Corpus">001F30</idno><idno type="wicri:explorRef" wicri:stream="PubMed" wicri:step="Corpus" wicri:corpus="PubMed">001F30</idno><idno type="wicri:Area/PubMed/Curation">001F06</idno><idno type="wicri:explorRef" wicri:stream="PubMed" wicri:step="Curation">001F06</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en">Application-Dedicated Selection of Filters (ADSF) using covariance maximization and orthogonal projection.</title><author><name sortKey="Hadoux, Xavier" sort="Hadoux, Xavier" uniqKey="Hadoux X" first="Xavier" last="Hadoux">Xavier Hadoux</name><affiliation wicri:level="1"><nlm:affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France; Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia. Electronic address: xavier@hadoux.com.</nlm:affiliation><country xml:lang="fr">Australie</country><wicri:regionArea>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France; Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery</wicri:regionArea></affiliation></author><author><name sortKey="Kumar, Dinesh Kant" sort="Kumar, Dinesh Kant" uniqKey="Kumar D" first="Dinesh Kant" last="Kumar">Dinesh Kant Kumar</name><affiliation wicri:level="1"><nlm:affiliation>Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001, Australia.</nlm:affiliation><country xml:lang="fr">Australie</country><wicri:regionArea>Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001</wicri:regionArea></affiliation></author><author><name sortKey="Sarossy, Marc G" sort="Sarossy, Marc G" uniqKey="Sarossy M" first="Marc G" last="Sarossy">Marc G. Sarossy</name><affiliation wicri:level="1"><nlm:affiliation>Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia; Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001, Australia.</nlm:affiliation><country xml:lang="fr">Australie</country><wicri:regionArea>Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia; Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001</wicri:regionArea></affiliation></author><author><name sortKey="Roger, Jean Michel" sort="Roger, Jean Michel" uniqKey="Roger J" first="Jean-Michel" last="Roger">Jean-Michel Roger</name><affiliation wicri:level="1"><nlm:affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France.</nlm:affiliation><country xml:lang="fr">France</country><wicri:regionArea>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5</wicri:regionArea></affiliation></author><author><name sortKey="Gorretta, Nathalie" sort="Gorretta, Nathalie" uniqKey="Gorretta N" first="Nathalie" last="Gorretta">Nathalie Gorretta</name><affiliation wicri:level="1"><nlm:affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France.</nlm:affiliation><country xml:lang="fr">France</country><wicri:regionArea>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5</wicri:regionArea></affiliation></author></analytic><series><title level="j">Analytica chimica acta</title><idno type="eISSN">1873-4324</idno><imprint><date when="2016" type="published">2016</date></imprint></series></biblStruct></sourceDesc></fileDesc><profileDesc><textClass></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">Visible and near-infrared (Vis-NIR) spectra are generated by the combination of numerous low resolution features. Spectral variables are thus highly correlated, which can cause problems for selecting the most appropriate ones for a given application. Some decomposition bases such as Fourier or wavelet generally help highlighting spectral features that are important, but are by nature constraint to have both positive and negative components. Thus, in addition to complicating the selected features interpretability, it impedes their use for application-dedicated sensors. In this paper we have proposed a new method for feature selection: Application-Dedicated Selection of Filters (ADSF). This method relaxes the shape constraint by enabling the selection of any type of user defined custom features. By considering only relevant features, based on the underlying nature of the data, high regularization of the final model can be obtained, even in the small sample size context often encountered in spectroscopic applications. For larger scale deployment of application-dedicated sensors, these predefined feature constraints can lead to application specific optical filters, e.g., lowpass, highpass, bandpass or bandstop filters with positive only coefficients. In a similar fashion to Partial Least Squares, ADSF successively selects features using covariance maximization and deflates their influences using orthogonal projection in order to optimally tune the selection to the data with limited redundancy. ADSF is well suited for spectroscopic data as it can deal with large numbers of highly correlated variables in supervised learning, even with many correlated responses.</div></front></TEI><pubmed><MedlineCitation Status="PubMed-not-MEDLINE" Owner="NLM"><PMID Version="1">27126785</PMID><DateCreated><Year>2016</Year><Month>04</Month><Day>29</Day></DateCreated><DateCompleted><Year>2017</Year><Month>06</Month><Day>13</Day></DateCompleted><DateRevised><Year>2017</Year><Month>08</Month><Day>01</Day></DateRevised><Article PubModel="Print-Electronic"><Journal><ISSN IssnType="Electronic">1873-4324</ISSN><JournalIssue CitedMedium="Internet"><Volume>921</Volume><PubDate><Year>2016</Year><Month>05</Month><Day>19</Day></PubDate></JournalIssue><Title>Analytica chimica acta</Title><ISOAbbreviation>Anal. Chim. Acta</ISOAbbreviation></Journal><ArticleTitle>Application-Dedicated Selection of Filters (ADSF) using covariance maximization and orthogonal projection.</ArticleTitle><Pagination><MedlinePgn>1-12</MedlinePgn></Pagination><ELocationID EIdType="pii" ValidYN="Y">S0003-2670(16)30426-3</ELocationID><ELocationID EIdType="doi" ValidYN="Y">10.1016/j.aca.2016.04.004</ELocationID><Abstract><AbstractText>Visible and near-infrared (Vis-NIR) spectra are generated by the combination of numerous low resolution features. Spectral variables are thus highly correlated, which can cause problems for selecting the most appropriate ones for a given application. Some decomposition bases such as Fourier or wavelet generally help highlighting spectral features that are important, but are by nature constraint to have both positive and negative components. Thus, in addition to complicating the selected features interpretability, it impedes their use for application-dedicated sensors. In this paper we have proposed a new method for feature selection: Application-Dedicated Selection of Filters (ADSF). This method relaxes the shape constraint by enabling the selection of any type of user defined custom features. By considering only relevant features, based on the underlying nature of the data, high regularization of the final model can be obtained, even in the small sample size context often encountered in spectroscopic applications. For larger scale deployment of application-dedicated sensors, these predefined feature constraints can lead to application specific optical filters, e.g., lowpass, highpass, bandpass or bandstop filters with positive only coefficients. In a similar fashion to Partial Least Squares, ADSF successively selects features using covariance maximization and deflates their influences using orthogonal projection in order to optimally tune the selection to the data with limited redundancy. ADSF is well suited for spectroscopic data as it can deal with large numbers of highly correlated variables in supervised learning, even with many correlated responses.</AbstractText><CopyrightInformation>Copyright © 2016 Elsevier B.V. All rights reserved.</CopyrightInformation></Abstract><AuthorList CompleteYN="Y"><Author ValidYN="Y"><LastName>Hadoux</LastName><ForeName>Xavier</ForeName><Initials>X</Initials><AffiliationInfo><Affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France; Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia. Electronic address: xavier@hadoux.com.</Affiliation></AffiliationInfo></Author><Author ValidYN="Y"><LastName>Kumar</LastName><ForeName>Dinesh Kant</ForeName><Initials>DK</Initials><AffiliationInfo><Affiliation>Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001, Australia.</Affiliation></AffiliationInfo></Author><Author ValidYN="Y"><LastName>Sarossy</LastName><ForeName>Marc G</ForeName><Initials>MG</Initials><AffiliationInfo><Affiliation>Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Australia; Ophthalmology, University of Melbourne, Department of Surgery, Australia; Biosignals Lab, School of Electrical and Computer Engineering, RMIT University, GPO Box 2476, Melbourne, VIC 3001, Australia.</Affiliation></AffiliationInfo></Author><Author ValidYN="Y"><LastName>Roger</LastName><ForeName>Jean-Michel</ForeName><Initials>JM</Initials><AffiliationInfo><Affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France.</Affiliation></AffiliationInfo></Author><Author ValidYN="Y"><LastName>Gorretta</LastName><ForeName>Nathalie</ForeName><Initials>N</Initials><AffiliationInfo><Affiliation>Irstea, UMR ITAP, 361 rue J-F Breton BP 5095, 34196 Montpellier Cedex 5, France.</Affiliation></AffiliationInfo></Author></AuthorList><Language>eng</Language><PublicationTypeList><PublicationType UI="D016428">Journal Article</PublicationType></PublicationTypeList><ArticleDate DateType="Electronic"><Year>2016</Year><Month>04</Month><Day>08</Day></ArticleDate></Article><MedlineJournalInfo><Country>Netherlands</Country><MedlineTA>Anal Chim Acta</MedlineTA><NlmUniqueID>0370534</NlmUniqueID><ISSNLinking>0003-2670</ISSNLinking></MedlineJournalInfo><KeywordList Owner="NOTNLM"><Keyword MajorTopicYN="Y">Dimension reduction</Keyword><Keyword MajorTopicYN="Y">Feature selection</Keyword><Keyword MajorTopicYN="Y">Filter selection</Keyword><Keyword MajorTopicYN="Y">Optical filters</Keyword><Keyword MajorTopicYN="Y">Orthogonal projection</Keyword><Keyword MajorTopicYN="Y">Variable selection</Keyword></KeywordList></MedlineCitation><PubmedData><History><PubMedPubDate PubStatus="received"><Year>2016</Year><Month>01</Month><Day>10</Day></PubMedPubDate><PubMedPubDate PubStatus="revised"><Year>2016</Year><Month>03</Month><Day>28</Day></PubMedPubDate><PubMedPubDate PubStatus="accepted"><Year>2016</Year><Month>04</Month><Day>05</Day></PubMedPubDate><PubMedPubDate PubStatus="entrez"><Year>2016</Year><Month>4</Month><Day>30</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate><PubMedPubDate PubStatus="pubmed"><Year>2016</Year><Month>4</Month><Day>30</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate><PubMedPubDate PubStatus="medline"><Year>2016</Year><Month>4</Month><Day>30</Day><Hour>6</Hour><Minute>1</Minute></PubMedPubDate></History><PublicationStatus>ppublish</PublicationStatus><ArticleIdList><ArticleId IdType="pubmed">27126785</ArticleId><ArticleId IdType="pii">S0003-2670(16)30426-3</ArticleId><ArticleId IdType="doi">10.1016/j.aca.2016.04.004</ArticleId></ArticleIdList></PubmedData></pubmed></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Asie/explor/AustralieFrV1/Data/PubMed/Curation
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 001F06 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/PubMed/Curation/biblio.hfd -nk 001F06 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Asie
|area= AustralieFrV1
|flux= PubMed
|étape= Curation
|type= RBID
|clé= pubmed:27126785
|texte= Application-Dedicated Selection of Filters (ADSF) using covariance maximization and orthogonal projection.
}}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/PubMed/Curation/RBID.i -Sk "pubmed:27126785" \
| HfdSelect -Kh $EXPLOR_AREA/Data/PubMed/Curation/biblio.hfd \
| NlmPubMed2Wicri -a AustralieFrV1
| This area was generated with Dilib version V0.6.33. |  |