Does the cost function matter in Bayes decision rule?
Identifieur interne : 000111 ( Ncbi/Checkpoint ); précédent : 000110; suivant : 000112Does the cost function matter in Bayes decision rule?
Auteurs : Ralf Schlü Ter [Allemagne] ; Markus Nussbaum-Thom ; Hermann NeySource :
- IEEE transactions on pattern analysis and machine intelligence [ 1939-3539 ] ; 2012.
English descriptors
- KwdEn :
- MESH :
Abstract
In many tasks in pattern recognition, such as automatic speech recognition (ASR), optical character recognition (OCR), part-of-speech (POS) tagging, and other string recognition tasks, we are faced with a well-known inconsistency: The Bayes decision rule is usually used to minimize string (symbol sequence) error, whereas, in practice, we want to minimize symbol (word, character, tag, etc.) error. When comparing different recognition systems, we do indeed use symbol error rate as an evaluation measure. The topic of this work is to analyze the relation between string (i.e., 0-1) and symbol error (i.e., metric, integer valued) cost functions in the Bayes decision rule, for which fundamental analytic results are derived. Simple conditions are derived for which the Bayes decision rule with integer-valued metric cost function and with 0-1 cost gives the same decisions or leads to classes with limited cost. The corresponding conditions can be tested with complexity linear in the number of classes. The results obtained do not make any assumption w.r.t. the structure of the underlying distributions or the classification problem. Nevertheless, the general analytic results are analyzed via simulations of string recognition problems with Levenshtein (edit) distance cost function. The results support earlier findings that considerable improvements are to be expected when initial error rates are high.
DOI: 10.1109/TPAMI.2011.163
PubMed: 21844628
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream PubMed, to step Corpus: 000034
- to stream PubMed, to step Curation: 000034
- to stream PubMed, to step Checkpoint: 000034
- to stream Ncbi, to step Merge: 000111
- to stream Ncbi, to step Curation: 000111
Links to Exploration step
pubmed:21844628Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">Does the cost function matter in Bayes decision rule?</title>
<author><name sortKey="Schlu Ter, Ralf" sort="Schlu Ter, Ralf" uniqKey="Schlu Ter R" first="Ralf" last="Schlü Ter">Ralf Schlü Ter</name>
<affiliation wicri:level="3"><nlm:affiliation>Lehrstuhl für Informatik 6, Computer Science Department, RWTH Aachen University, Ahornstr. 55, Aachen 52074, Germany. schlueter@cs.rwth-aachen.de</nlm:affiliation>
<country xml:lang="fr">Allemagne</country>
<wicri:regionArea>Lehrstuhl für Informatik 6, Computer Science Department, RWTH Aachen University, Ahornstr. 55, Aachen 52074</wicri:regionArea>
<placeName><region type="land" nuts="1">Rhénanie-du-Nord-Westphalie</region>
<region type="district" nuts="2">District de Cologne</region>
<settlement type="city">Aix-la-Chapelle</settlement>
</placeName>
</affiliation>
</author>
<author><name sortKey="Nussbaum Thom, Markus" sort="Nussbaum Thom, Markus" uniqKey="Nussbaum Thom M" first="Markus" last="Nussbaum-Thom">Markus Nussbaum-Thom</name>
</author>
<author><name sortKey="Ney, Hermann" sort="Ney, Hermann" uniqKey="Ney H" first="Hermann" last="Ney">Hermann Ney</name>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">PubMed</idno>
<date when="2012">2012</date>
<idno type="doi">10.1109/TPAMI.2011.163</idno>
<idno type="RBID">pubmed:21844628</idno>
<idno type="pmid">21844628</idno>
<idno type="wicri:Area/PubMed/Corpus">000034</idno>
<idno type="wicri:Area/PubMed/Curation">000034</idno>
<idno type="wicri:Area/PubMed/Checkpoint">000034</idno>
<idno type="wicri:Area/Ncbi/Merge">000111</idno>
<idno type="wicri:Area/Ncbi/Curation">000111</idno>
<idno type="wicri:Area/Ncbi/Checkpoint">000111</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en">Does the cost function matter in Bayes decision rule?</title>
<author><name sortKey="Schlu Ter, Ralf" sort="Schlu Ter, Ralf" uniqKey="Schlu Ter R" first="Ralf" last="Schlü Ter">Ralf Schlü Ter</name>
<affiliation wicri:level="3"><nlm:affiliation>Lehrstuhl für Informatik 6, Computer Science Department, RWTH Aachen University, Ahornstr. 55, Aachen 52074, Germany. schlueter@cs.rwth-aachen.de</nlm:affiliation>
<country xml:lang="fr">Allemagne</country>
<wicri:regionArea>Lehrstuhl für Informatik 6, Computer Science Department, RWTH Aachen University, Ahornstr. 55, Aachen 52074</wicri:regionArea>
<placeName><region type="land" nuts="1">Rhénanie-du-Nord-Westphalie</region>
<region type="district" nuts="2">District de Cologne</region>
<settlement type="city">Aix-la-Chapelle</settlement>
</placeName>
</affiliation>
</author>
<author><name sortKey="Nussbaum Thom, Markus" sort="Nussbaum Thom, Markus" uniqKey="Nussbaum Thom M" first="Markus" last="Nussbaum-Thom">Markus Nussbaum-Thom</name>
</author>
<author><name sortKey="Ney, Hermann" sort="Ney, Hermann" uniqKey="Ney H" first="Hermann" last="Ney">Hermann Ney</name>
</author>
</analytic>
<series><title level="j">IEEE transactions on pattern analysis and machine intelligence</title>
<idno type="eISSN">1939-3539</idno>
<imprint><date when="2012" type="published">2012</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Algorithms</term>
<term>Bayes Theorem</term>
<term>Computer Simulation</term>
<term>Pattern Recognition, Automated (methods)</term>
<term>Speech Recognition Software</term>
</keywords>
<keywords scheme="MESH" qualifier="methods" xml:lang="en"><term>Pattern Recognition, Automated</term>
</keywords>
<keywords scheme="MESH" xml:lang="en"><term>Algorithms</term>
<term>Bayes Theorem</term>
<term>Computer Simulation</term>
<term>Speech Recognition Software</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">In many tasks in pattern recognition, such as automatic speech recognition (ASR), optical character recognition (OCR), part-of-speech (POS) tagging, and other string recognition tasks, we are faced with a well-known inconsistency: The Bayes decision rule is usually used to minimize string (symbol sequence) error, whereas, in practice, we want to minimize symbol (word, character, tag, etc.) error. When comparing different recognition systems, we do indeed use symbol error rate as an evaluation measure. The topic of this work is to analyze the relation between string (i.e., 0-1) and symbol error (i.e., metric, integer valued) cost functions in the Bayes decision rule, for which fundamental analytic results are derived. Simple conditions are derived for which the Bayes decision rule with integer-valued metric cost function and with 0-1 cost gives the same decisions or leads to classes with limited cost. The corresponding conditions can be tested with complexity linear in the number of classes. The results obtained do not make any assumption w.r.t. the structure of the underlying distributions or the classification problem. Nevertheless, the general analytic results are analyzed via simulations of string recognition problems with Levenshtein (edit) distance cost function. The results support earlier findings that considerable improvements are to be expected when initial error rates are high.</div>
</front>
</TEI>
<affiliations><list><country><li>Allemagne</li>
</country>
<region><li>District de Cologne</li>
<li>Rhénanie-du-Nord-Westphalie</li>
</region>
<settlement><li>Aix-la-Chapelle</li>
</settlement>
</list>
<tree><noCountry><name sortKey="Ney, Hermann" sort="Ney, Hermann" uniqKey="Ney H" first="Hermann" last="Ney">Hermann Ney</name>
<name sortKey="Nussbaum Thom, Markus" sort="Nussbaum Thom, Markus" uniqKey="Nussbaum Thom M" first="Markus" last="Nussbaum-Thom">Markus Nussbaum-Thom</name>
</noCountry>
<country name="Allemagne"><region name="Rhénanie-du-Nord-Westphalie"><name sortKey="Schlu Ter, Ralf" sort="Schlu Ter, Ralf" uniqKey="Schlu Ter R" first="Ralf" last="Schlü Ter">Ralf Schlü Ter</name>
</region>
</country>
</tree>
</affiliations>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/Ncbi/Checkpoint
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000111 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Ncbi/Checkpoint/biblio.hfd -nk 000111 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Ticri/CIDE |area= OcrV1 |flux= Ncbi |étape= Checkpoint |type= RBID |clé= pubmed:21844628 |texte= Does the cost function matter in Bayes decision rule? }}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/Ncbi/Checkpoint/RBID.i -Sk "pubmed:21844628" \ | HfdSelect -Kh $EXPLOR_AREA/Data/Ncbi/Checkpoint/biblio.hfd \ | NlmPubMed2Wicri -a OcrV1
This area was generated with Dilib version V0.6.32. |