Apportioned margin approach for cost sensitive large margin classifiers View Full Text


Ontology type: schema:ScholarlyArticle      Open Access: True


Article Info

DATE

2021-10-08

AUTHORS

Lee-Ad Gottlieb, Eran Kaufman, Aryeh Kontorovich

ABSTRACT

We consider the problem of cost sensitive multiclass classification, where we would like to increase the sensitivity of an important class at the expense of a less important one. We adopt an apportioned margin framework to address this problem, which enables an efficient margin shift between classes that share the same boundary. The decision boundary between all pairs of classes divides the margin between them in accordance with a given prioritization vector, which yields a tighter error bound for the important classes while also reducing the overall out-of-sample error. In addition to demonstrating an efficient implementation of our framework, we derive generalization bounds, demonstrate Fisher consistency, adapt the framework to Mercer’s kernel and to neural networks, and report promising empirical results on all accounts. More... »

PAGES

1215-1235

References to SciGraph publications

  • 1991. Probability in Banach Spaces, Isoperimetry and Processes in NONE
  • 2003. Support Vector Machines with Example Dependent Costs in MACHINE LEARNING: ECML 2003
  • 2001-09-13. A Generalized Representer Theorem in COMPUTATIONAL LEARNING THEORY
  • 1998-02. Machine Learning for the Detection of Oil Spills in Satellite Radar Images in MACHINE LEARNING
  • 2005-05. Multicategory Proximal Support Vector Machine Classifiers in MACHINE LEARNING
  • 1974. Theoretical Statistics in NONE
  • 1999-01. Multicategory Classification by Support Vector Machines in COMPUTATIONAL OPTIMIZATION AND APPLICATIONS
  • 2011. Feature and Sample Weighted Support Vector Machine in KNOWLEDGE ENGINEERING AND MANAGEMENT
  • 1997-09. Adaptive Fraud Detection in DATA MINING AND KNOWLEDGE DISCOVERY
  • 2018. Learning from Imbalanced Data Sets in NONE
  • 2010-10-16. Pegasos: primal estimated sub-gradient solver for SVM in MATHEMATICAL PROGRAMMING
  • Identifiers

    URI

    http://scigraph.springernature.com/pub.10.1007/s10472-021-09776-w

    DOI

    http://dx.doi.org/10.1007/s10472-021-09776-w

    DIMENSIONS

    https://app.dimensions.ai/details/publication/pub.1141747401


    Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
    Incoming Citations Browse incoming citations for this publication using opencitations.net

    JSON-LD is the canonical representation for SciGraph data.

    TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

    [
      {
        "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
        "about": [
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Information and Computing Sciences", 
            "type": "DefinedTerm"
          }, 
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Artificial Intelligence and Image Processing", 
            "type": "DefinedTerm"
          }
        ], 
        "author": [
          {
            "affiliation": {
              "alternateName": "Ariel University, Ariel, Israel", 
              "id": "http://www.grid.ac/institutes/grid.411434.7", 
              "name": [
                "Ariel University, Ariel, Israel"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Gottlieb", 
            "givenName": "Lee-Ad", 
            "id": "sg:person.016113030607.10", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.016113030607.10"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Ben-Gurion University, Beer Sheva, Israel", 
              "id": "http://www.grid.ac/institutes/grid.7489.2", 
              "name": [
                "Ben-Gurion University, Beer Sheva, Israel"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Kaufman", 
            "givenName": "Eran", 
            "id": "sg:person.013375351761.01", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013375351761.01"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Ben-Gurion University, Beer Sheva, Israel", 
              "id": "http://www.grid.ac/institutes/grid.7489.2", 
              "name": [
                "Ben-Gurion University, Beer Sheva, Israel"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Kontorovich", 
            "givenName": "Aryeh", 
            "id": "sg:person.0770077045.42", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.0770077045.42"
            ], 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "sg:pub.10.1007/s10107-010-0420-4", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1050541291", 
              "https://doi.org/10.1007/s10107-010-0420-4"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/s10994-005-0463-6", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1024280650", 
              "https://doi.org/10.1007/s10994-005-0463-6"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1023/a:1009700419189", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1001263423", 
              "https://doi.org/10.1023/a:1009700419189"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-319-98074-4", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1108232028", 
              "https://doi.org/10.1007/978-3-319-98074-4"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-25661-5_47", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1028533169", 
              "https://doi.org/10.1007/978-3-642-25661-5_47"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-540-39857-8_5", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1007099244", 
              "https://doi.org/10.1007/978-3-540-39857-8_5"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-1-4899-2887-0", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1109705820", 
              "https://doi.org/10.1007/978-1-4899-2887-0"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1023/a:1008663629662", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1038382810", 
              "https://doi.org/10.1023/a:1008663629662"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-20212-4", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1008285500", 
              "https://doi.org/10.1007/978-3-642-20212-4"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/3-540-44581-1_27", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1019980111", 
              "https://doi.org/10.1007/3-540-44581-1_27"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1023/a:1007452223027", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1025850532", 
              "https://doi.org/10.1023/a:1007452223027"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "2021-10-08", 
        "datePublishedReg": "2021-10-08", 
        "description": "We consider the problem of cost sensitive multiclass classification, where we would like to increase the sensitivity of an important class at the expense of a less important one. We adopt an apportioned margin framework to address this problem, which enables an efficient margin shift between classes that share the same boundary. The decision boundary between all pairs of classes divides the margin between them in accordance with a given prioritization vector, which yields a tighter error bound for the important classes while also reducing the overall out-of-sample error. In addition to demonstrating an efficient implementation of our framework, we derive generalization bounds, demonstrate Fisher consistency, adapt the framework to Mercer\u2019s kernel and to neural networks, and report promising empirical results on all accounts.", 
        "genre": "article", 
        "id": "sg:pub.10.1007/s10472-021-09776-w", 
        "inLanguage": "en", 
        "isAccessibleForFree": true, 
        "isPartOf": [
          {
            "id": "sg:journal.1043955", 
            "issn": [
              "1012-2443", 
              "1573-7470"
            ], 
            "name": "Annals of Mathematics and Artificial Intelligence", 
            "publisher": "Springer Nature", 
            "type": "Periodical"
          }, 
          {
            "issueNumber": "12", 
            "type": "PublicationIssue"
          }, 
          {
            "type": "PublicationVolume", 
            "volumeNumber": "89"
          }
        ], 
        "keywords": [
          "promising empirical results", 
          "large margin classifiers", 
          "multiclass classification", 
          "neural network", 
          "margin classifier", 
          "efficient implementation", 
          "margin framework", 
          "pair of classes", 
          "generalization bounds", 
          "Mercer kernel", 
          "tight error", 
          "margin approach", 
          "important class", 
          "framework", 
          "kernel", 
          "empirical results", 
          "classifier", 
          "Fisher consistency", 
          "sample error", 
          "network", 
          "important one", 
          "error", 
          "implementation", 
          "classification", 
          "class", 
          "decisions", 
          "bounds", 
          "vector", 
          "consistency", 
          "one", 
          "expense", 
          "results", 
          "pairs", 
          "account", 
          "boundaries", 
          "accordance", 
          "addition", 
          "same boundary", 
          "margin", 
          "shift", 
          "sensitivity", 
          "problem", 
          "approach", 
          "cost sensitive multiclass classification", 
          "sensitive multiclass classification", 
          "apportioned margin framework", 
          "efficient margin shift", 
          "margin shift", 
          "prioritization vector", 
          "Apportioned margin approach", 
          "cost sensitive large margin classifiers", 
          "sensitive large margin classifiers"
        ], 
        "name": "Apportioned margin approach for cost sensitive large margin classifiers", 
        "pagination": "1215-1235", 
        "productId": [
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "pub.1141747401"
            ]
          }, 
          {
            "name": "doi", 
            "type": "PropertyValue", 
            "value": [
              "10.1007/s10472-021-09776-w"
            ]
          }
        ], 
        "sameAs": [
          "https://doi.org/10.1007/s10472-021-09776-w", 
          "https://app.dimensions.ai/details/publication/pub.1141747401"
        ], 
        "sdDataset": "articles", 
        "sdDatePublished": "2022-01-01T19:02", 
        "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
        "sdPublisher": {
          "name": "Springer Nature - SN SciGraph project", 
          "type": "Organization"
        }, 
        "sdSource": "s3://com-springernature-scigraph/baseset/20220101/entities/gbq_results/article/article_909.jsonl", 
        "type": "ScholarlyArticle", 
        "url": "https://doi.org/10.1007/s10472-021-09776-w"
      }
    ]
     

    Download the RDF metadata as:  json-ld nt turtle xml License info

    HOW TO GET THIS DATA PROGRAMMATICALLY:

    JSON-LD is a popular format for linked data which is fully compatible with JSON.

    curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/s10472-021-09776-w'

    N-Triples is a line-based linked data format ideal for batch operations.

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/s10472-021-09776-w'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/s10472-021-09776-w'

    RDF/XML is a standard XML format for linked data.

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/s10472-021-09776-w'


     

    This table displays all metadata directly associated to this object as RDF triples.

    171 TRIPLES      22 PREDICATES      88 URIs      69 LITERALS      6 BLANK NODES

    Subject Predicate Object
    1 sg:pub.10.1007/s10472-021-09776-w schema:about anzsrc-for:08
    2 anzsrc-for:0801
    3 schema:author N9814b75608ac4c0f83304ff451e23cf5
    4 schema:citation sg:pub.10.1007/3-540-44581-1_27
    5 sg:pub.10.1007/978-1-4899-2887-0
    6 sg:pub.10.1007/978-3-319-98074-4
    7 sg:pub.10.1007/978-3-540-39857-8_5
    8 sg:pub.10.1007/978-3-642-20212-4
    9 sg:pub.10.1007/978-3-642-25661-5_47
    10 sg:pub.10.1007/s10107-010-0420-4
    11 sg:pub.10.1007/s10994-005-0463-6
    12 sg:pub.10.1023/a:1007452223027
    13 sg:pub.10.1023/a:1008663629662
    14 sg:pub.10.1023/a:1009700419189
    15 schema:datePublished 2021-10-08
    16 schema:datePublishedReg 2021-10-08
    17 schema:description We consider the problem of cost sensitive multiclass classification, where we would like to increase the sensitivity of an important class at the expense of a less important one. We adopt an apportioned margin framework to address this problem, which enables an efficient margin shift between classes that share the same boundary. The decision boundary between all pairs of classes divides the margin between them in accordance with a given prioritization vector, which yields a tighter error bound for the important classes while also reducing the overall out-of-sample error. In addition to demonstrating an efficient implementation of our framework, we derive generalization bounds, demonstrate Fisher consistency, adapt the framework to Mercer’s kernel and to neural networks, and report promising empirical results on all accounts.
    18 schema:genre article
    19 schema:inLanguage en
    20 schema:isAccessibleForFree true
    21 schema:isPartOf N31e730035cd84a3d87ed5b1053c4fc9c
    22 N433e499f1c55447e92d1a97410aaa99e
    23 sg:journal.1043955
    24 schema:keywords Apportioned margin approach
    25 Fisher consistency
    26 Mercer kernel
    27 accordance
    28 account
    29 addition
    30 apportioned margin framework
    31 approach
    32 boundaries
    33 bounds
    34 class
    35 classification
    36 classifier
    37 consistency
    38 cost sensitive large margin classifiers
    39 cost sensitive multiclass classification
    40 decisions
    41 efficient implementation
    42 efficient margin shift
    43 empirical results
    44 error
    45 expense
    46 framework
    47 generalization bounds
    48 implementation
    49 important class
    50 important one
    51 kernel
    52 large margin classifiers
    53 margin
    54 margin approach
    55 margin classifier
    56 margin framework
    57 margin shift
    58 multiclass classification
    59 network
    60 neural network
    61 one
    62 pair of classes
    63 pairs
    64 prioritization vector
    65 problem
    66 promising empirical results
    67 results
    68 same boundary
    69 sample error
    70 sensitive large margin classifiers
    71 sensitive multiclass classification
    72 sensitivity
    73 shift
    74 tight error
    75 vector
    76 schema:name Apportioned margin approach for cost sensitive large margin classifiers
    77 schema:pagination 1215-1235
    78 schema:productId N1ff1c81b9dd0412890a8ede176010e70
    79 Nbf51ae53198744b4bc9622fd44c28c41
    80 schema:sameAs https://app.dimensions.ai/details/publication/pub.1141747401
    81 https://doi.org/10.1007/s10472-021-09776-w
    82 schema:sdDatePublished 2022-01-01T19:02
    83 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    84 schema:sdPublisher N91ec73a0f9fd46d890dbe6ceda7381f0
    85 schema:url https://doi.org/10.1007/s10472-021-09776-w
    86 sgo:license sg:explorer/license/
    87 sgo:sdDataset articles
    88 rdf:type schema:ScholarlyArticle
    89 N1ff1c81b9dd0412890a8ede176010e70 schema:name doi
    90 schema:value 10.1007/s10472-021-09776-w
    91 rdf:type schema:PropertyValue
    92 N31e730035cd84a3d87ed5b1053c4fc9c schema:issueNumber 12
    93 rdf:type schema:PublicationIssue
    94 N433e499f1c55447e92d1a97410aaa99e schema:volumeNumber 89
    95 rdf:type schema:PublicationVolume
    96 N83807919ad584380aa6da70cc076b4df rdf:first sg:person.013375351761.01
    97 rdf:rest Naa429168d4664117a0fdcca735c1540c
    98 N91ec73a0f9fd46d890dbe6ceda7381f0 schema:name Springer Nature - SN SciGraph project
    99 rdf:type schema:Organization
    100 N9814b75608ac4c0f83304ff451e23cf5 rdf:first sg:person.016113030607.10
    101 rdf:rest N83807919ad584380aa6da70cc076b4df
    102 Naa429168d4664117a0fdcca735c1540c rdf:first sg:person.0770077045.42
    103 rdf:rest rdf:nil
    104 Nbf51ae53198744b4bc9622fd44c28c41 schema:name dimensions_id
    105 schema:value pub.1141747401
    106 rdf:type schema:PropertyValue
    107 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
    108 schema:name Information and Computing Sciences
    109 rdf:type schema:DefinedTerm
    110 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
    111 schema:name Artificial Intelligence and Image Processing
    112 rdf:type schema:DefinedTerm
    113 sg:journal.1043955 schema:issn 1012-2443
    114 1573-7470
    115 schema:name Annals of Mathematics and Artificial Intelligence
    116 schema:publisher Springer Nature
    117 rdf:type schema:Periodical
    118 sg:person.013375351761.01 schema:affiliation grid-institutes:grid.7489.2
    119 schema:familyName Kaufman
    120 schema:givenName Eran
    121 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013375351761.01
    122 rdf:type schema:Person
    123 sg:person.016113030607.10 schema:affiliation grid-institutes:grid.411434.7
    124 schema:familyName Gottlieb
    125 schema:givenName Lee-Ad
    126 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.016113030607.10
    127 rdf:type schema:Person
    128 sg:person.0770077045.42 schema:affiliation grid-institutes:grid.7489.2
    129 schema:familyName Kontorovich
    130 schema:givenName Aryeh
    131 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.0770077045.42
    132 rdf:type schema:Person
    133 sg:pub.10.1007/3-540-44581-1_27 schema:sameAs https://app.dimensions.ai/details/publication/pub.1019980111
    134 https://doi.org/10.1007/3-540-44581-1_27
    135 rdf:type schema:CreativeWork
    136 sg:pub.10.1007/978-1-4899-2887-0 schema:sameAs https://app.dimensions.ai/details/publication/pub.1109705820
    137 https://doi.org/10.1007/978-1-4899-2887-0
    138 rdf:type schema:CreativeWork
    139 sg:pub.10.1007/978-3-319-98074-4 schema:sameAs https://app.dimensions.ai/details/publication/pub.1108232028
    140 https://doi.org/10.1007/978-3-319-98074-4
    141 rdf:type schema:CreativeWork
    142 sg:pub.10.1007/978-3-540-39857-8_5 schema:sameAs https://app.dimensions.ai/details/publication/pub.1007099244
    143 https://doi.org/10.1007/978-3-540-39857-8_5
    144 rdf:type schema:CreativeWork
    145 sg:pub.10.1007/978-3-642-20212-4 schema:sameAs https://app.dimensions.ai/details/publication/pub.1008285500
    146 https://doi.org/10.1007/978-3-642-20212-4
    147 rdf:type schema:CreativeWork
    148 sg:pub.10.1007/978-3-642-25661-5_47 schema:sameAs https://app.dimensions.ai/details/publication/pub.1028533169
    149 https://doi.org/10.1007/978-3-642-25661-5_47
    150 rdf:type schema:CreativeWork
    151 sg:pub.10.1007/s10107-010-0420-4 schema:sameAs https://app.dimensions.ai/details/publication/pub.1050541291
    152 https://doi.org/10.1007/s10107-010-0420-4
    153 rdf:type schema:CreativeWork
    154 sg:pub.10.1007/s10994-005-0463-6 schema:sameAs https://app.dimensions.ai/details/publication/pub.1024280650
    155 https://doi.org/10.1007/s10994-005-0463-6
    156 rdf:type schema:CreativeWork
    157 sg:pub.10.1023/a:1007452223027 schema:sameAs https://app.dimensions.ai/details/publication/pub.1025850532
    158 https://doi.org/10.1023/a:1007452223027
    159 rdf:type schema:CreativeWork
    160 sg:pub.10.1023/a:1008663629662 schema:sameAs https://app.dimensions.ai/details/publication/pub.1038382810
    161 https://doi.org/10.1023/a:1008663629662
    162 rdf:type schema:CreativeWork
    163 sg:pub.10.1023/a:1009700419189 schema:sameAs https://app.dimensions.ai/details/publication/pub.1001263423
    164 https://doi.org/10.1023/a:1009700419189
    165 rdf:type schema:CreativeWork
    166 grid-institutes:grid.411434.7 schema:alternateName Ariel University, Ariel, Israel
    167 schema:name Ariel University, Ariel, Israel
    168 rdf:type schema:Organization
    169 grid-institutes:grid.7489.2 schema:alternateName Ben-Gurion University, Beer Sheva, Israel
    170 schema:name Ben-Gurion University, Beer Sheva, Israel
    171 rdf:type schema:Organization
     




    Preview window. Press ESC to close (or click here)


    ...