TS-MULE : Local Interpretable Model-Agnostic Explanations for Time Series Forecast Models

Schlegel, Udo; Vo, Duy Lam; Keim, Daniel A.; Seebacher, Daniel

TS-MULE : Local Interpretable Model-Agnostic Explanations for Time Series Forecast Models

Files

Schlegel_2-vxq3bhqko6i12.pdf(663.51 KB)

Date

2022

Publication type

Contribution to a conference collection

Publication status

Published

Published in

Machine Learning and Principles and Practice of Knowledge Discovery in Databases : International Workshops of ECML PKDD 2021, Proceedings, Part I / Kamp, Michael et al. (ed.). - Cham : Springer, 2022. - (Communications in Computer and Information Science ; 1524). - pp. 5-14. - ISBN 978-3-030-93735-5

Abstract

Time series forecasting is a demanding task ranging from weather to failure forecasting with black-box models achieving state-of-the-art performances. However, understanding and debugging are not guaranteed. We propose TS-MULE, a local surrogate model explanation method specialized for time series extending the LIME approach. Our extended LIME works with various ways to segment and perturb the time series data. In our extension, we present six sampling segmentation approaches for time series to improve the quality of surrogate attributions and demonstrate their performances on three deep learning model architectures and three common multivariate time series datasets.

Subject (DDC)

004 Computer Science

Conference

Machine Learning and Principles and Practice of Knowledge Discovery in Databases : International Workshops of ECML PKDD 2021, Sep 13, 2021 - Sep 17, 2021

Cite This

ISO 690

SCHLEGEL, Udo, Duy Lam VO, Daniel A. KEIM, Daniel SEEBACHER, 2022. TS-MULE : Local Interpretable Model-Agnostic Explanations for Time Series Forecast Models. Machine Learning and Principles and Practice of Knowledge Discovery in Databases : International Workshops of ECML PKDD 2021, Sep 13, 2021 - Sep 17, 2021. In: KAMP, Michael, ed. and others. Machine Learning and Principles and Practice of Knowledge Discovery in Databases : International Workshops of ECML PKDD 2021, Proceedings, Part I. Cham:Springer, pp. 5-14. ISBN 978-3-030-93735-5. Available under: doi: 10.1007/978-3-030-93736-2_1

BibTex

@inproceedings{Schlegel2022TSMUL-55037,
  year={2022},
  doi={10.1007/978-3-030-93736-2_1},
  title={TS-MULE : Local Interpretable Model-Agnostic Explanations for Time Series Forecast Models},
  number={1524},
  isbn={978-3-030-93735-5},
  publisher={Springer},
  address={Cham},
  series={Communications in Computer and Information Science},
  booktitle={Machine Learning and Principles and Practice of Knowledge Discovery in Databases : International Workshops of ECML PKDD 2021, Proceedings, Part I},
  pages={5--14},
  editor={Kamp, Michael},
  author={Schlegel, Udo and Vo, Duy Lam and Keim, Daniel A. and Seebacher, Daniel}
}

RDF

<rdf:RDF
    xmlns:dcterms="http://purl.org/dc/terms/"
    xmlns:dc="http://purl.org/dc/elements/1.1/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:bibo="http://purl.org/ontology/bibo/"
    xmlns:dspace="http://digital-repositories.org/ontologies/dspace/0.1.0#"
    xmlns:foaf="http://xmlns.com/foaf/0.1/"
    xmlns:void="http://rdfs.org/ns/void#"
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" > 
  <rdf:Description rdf:about="https://kops.uni-konstanz.de/server/rdf/resource/123456789/55037">
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/55037"/>
    <dc:creator>Vo, Duy Lam</dc:creator>
    <dc:language>eng</dc:language>
    <dcterms:abstract xml:lang="eng">Time series forecasting is a demanding task ranging from weather to failure forecasting with black-box models achieving state-of-the-art performances. However, understanding and debugging are not guaranteed. We propose TS-MULE, a local surrogate model explanation method specialized for time series extending the LIME approach. Our extended LIME works with various ways to segment and perturb the time series data. In our extension, we present six sampling segmentation approaches for time series to improve the quality of surrogate attributions and demonstrate their performances on three deep learning model architectures and three common multivariate time series datasets.</dcterms:abstract>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-09-27T13:05:59Z</dc:date>
    <dc:rights>terms-of-use</dc:rights>
    <dcterms:issued>2022</dcterms:issued>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:creator>Schlegel, Udo</dc:creator>
    <dc:contributor>Seebacher, Daniel</dc:contributor>
    <dspace:hasBitstream rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/55037/1/Schlegel_2-vxq3bhqko6i12.pdf"/>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dcterms:title>TS-MULE : Local Interpretable Model-Agnostic Explanations for Time Series Forecast Models</dcterms:title>
    <dc:contributor>Schlegel, Udo</dc:contributor>
    <dc:contributor>Keim, Daniel A.</dc:contributor>
    <dc:creator>Keim, Daniel A.</dc:creator>
    <dcterms:hasPart rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/55037/1/Schlegel_2-vxq3bhqko6i12.pdf"/>
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-09-27T13:05:59Z</dcterms:available>
    <dcterms:rights rdf:resource="https://rightsstatements.org/page/InC/1.0/"/>
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
    <dc:creator>Seebacher, Daniel</dc:creator>
    <dc:contributor>Vo, Duy Lam</dc:contributor>
  </rdf:Description>
</rdf:RDF>

Bibliography of Konstanz

Yes