Analyzing and repairing concept drift adaptation in data stream classification

Simple item page
dc.contributor.authorHalstead, Benen_NZ
dc.contributor.authorKoh, Yun Singen_NZ
dc.contributor.authorRiddle, Patriciaen_NZ
dc.contributor.authorPears, Russelen_NZ
dc.contributor.authorPechenizkiy, Mykolaen_NZ
dc.contributor.authorBifet, Alberten_NZ
dc.contributor.authorOlivares, Gustavoen_NZ
dc.contributor.authorCoulson, Guyen_NZ
dc.date.accessioned2022-08-21T09:28:08Z
dc.date.available2022-08-21T09:28:08Z
dc.date.issued2021en_NZ
dc.description.abstractData collected over time often exhibit changes in distribution, or concept drift, caused by changes in factors relevant to the classification task, e.g. weather conditions. Incorporating all relevant factors into the model may be able to capture these changes, however, this is usually not practical. Data stream based methods, which instead explicitly detect concept drift, have been shown to retain performance under unknown changing conditions. These methods adapt to concept drift by training a model to classify each distinct data distribution. However, we hypothesize that existing methods do not robustly handle real-world tasks, leading to adaptation errors where context is misidentified. Adaptation errors may cause a system to use a model which does not fit the current data, reducing performance. We propose a novel repair algorithm to identify and correct errors in concept drift adaptation. Evaluation on synthetic data shows that our proposed AiRStream system has higher performance than baseline methods, while is also better at capturing the dynamics of the stream. Evaluation on an air quality inference task shows AiRStream provides increased real-world performance compared to eight baseline methods. A case study shows that AiRStream is able to build a robust model of environmental conditions over this task, allowing the adaptions made to concept drift to be analysed and related to changes in weather. We discovered a strong predictive link between the adaptions made by AiRStream and changes in meteorological conditions.en_NZ
dc.format.mimetypeapplication/pdf
dc.identifier.doi10.1007/s10994-021-05993-wen_NZ
dc.identifier.eissn1573-0565en_NZ
dc.identifier.issn0885-6125en_NZ
dc.identifier.urihttps://hdl.handle.net/10289/15050
dc.language.isoen
dc.relation.isPartOfMachine Learningen_NZ
dc.rights© The Author(s), under exclusive licence to Springer Science+Business Media LLC, part of Springer Nature 2021
dc.subjectcomputer scienceen_NZ
dc.subjectconcept driften_NZ
dc.subjectdata stream classificationen_NZ
dc.subjectrecurring conceptsen_NZ
dc.titleAnalyzing and repairing concept drift adaptation in data stream classificationen_NZ
dc.typeJournal Article
pubs.elements-id261629
pubs.organisational-group/Waikato
pubs.organisational-group/Waikato/2026 PBRF
pubs.organisational-group/Waikato/DHECS
pubs.organisational-group/Waikato/DHECS/2026 PBRF - DHEC
pubs.organisational-group/Waikato/DHECS/ARII
pubs.organisational-group/Waikato/DHECS/ARII/2026 PBRF - ARII
pubs.publication-statusPublisheden_NZ
pubs.user.infoBifet, Albert (albert.bifetfiguerol@waikato.ac.nz)
uow.verification.statusverified
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Halstead2021_Article_AnalyzingAndRepairingConceptDr.pdf
Size:
1.95 MB
Format:
Adobe Portable Document Format
Description:
Published version
Thumbnail Image
Name: Halstead2021_Article_AnalyzingAndRepairingConceptDr.pdf
Size: 1.95 MB
Kind: Adobe PDF
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
Research Commons Deposit Agreement 2017.pdf
Size:
188.11 KB
Format:
Adobe Portable Document Format
Description:
Name: Research Commons Deposit Agreement 2017.pdf
Size: 188.11 KB
Kind: Adobe PDF
Collections
Computing and Mathematical Sciences Papers