Show simple item record  

dc.contributor.authorGomes, Heitor Muriloen_NZ
dc.contributor.authorBifet, Alberten_NZ
dc.contributor.authorRead, Jesseen_NZ
dc.contributor.authorBarddal, Jean Paulen_NZ
dc.contributor.authorEnembreck, Fabrícioen_NZ
dc.contributor.authorPfahringer, Bernharden_NZ
dc.contributor.authorHolmes, Geoffreyen_NZ
dc.contributor.authorAbdessalem, Talelen_NZ
dc.date.accessioned2017-07-26T21:51:48Z
dc.date.available2017en_NZ
dc.date.available2017-07-26T21:51:48Z
dc.date.issued2017en_NZ
dc.identifier.citationGomes, H. M., Bifet, A., Read, J., Barddal, J. P., Enembreck, F., Pfahringer, B., … Abdessalem, T. (2017). Adaptive random forests for evolving data stream classification. Machine Learning, (Online First), 1–27. https://doi.org/10.1007/s10994-017-5642-8en
dc.identifier.issn0885-6125en_NZ
dc.identifier.urihttps://hdl.handle.net/10289/11231
dc.description.abstractRandom forests is currently one of the most used machine learning algorithms in the non-streaming (batch) setting. This preference is attributable to its high learning performance and low demands with respect to input preparation and hyper-parameter tuning. However, in the challenging context of evolving data streams, there is no random forests algorithm that can be considered state-of-the-art in comparison to bagging and boosting based algorithms. In this work, we present the adaptive random forest (ARF) algorithm for classification of evolving data streams. In contrast to previous attempts of replicating random forests for data stream learning, ARF includes an effective resampling method and adaptive operators that can cope with different types of concept drifts without complex optimizations for different data sets. We present experiments with a parallel implementation of ARF which has no degradation in terms of classification performance in comparison to a serial implementation, since trees and adaptive operators are independent from one another. Finally, we compare ARF with state-of-the-art algorithms in a traditional test-then-train evaluation and a novel delayed labelling evaluation, and show that ARF is accurate and uses a feasible amount of resources.
dc.format.mimetypeapplication/pdf
dc.language.isoen
dc.publisherSpringeren_NZ
dc.rights© 2017 Springer International Publishing Switzerland.This is the author's accepted version. The final publication is available at Springer via dx.doi.org/10.1007/s10994-017-5642-8
dc.subjectData stream mining
dc.subjectRandom forests
dc.subjectEnsemble learning
dc.subjectConcept drift
dc.subjectMachine learning
dc.titleAdaptive random forests for evolving data stream classificationen_NZ
dc.typeJournal Article
dc.identifier.doi10.1007/s10994-017-5642-8en_NZ
dc.relation.isPartOfMachine learningen_NZ
pubs.begin-page1
pubs.elements-id195129
pubs.end-page27
pubs.issueOnline Firsten_NZ
pubs.publication-statusPublished onlineen_NZ
dc.identifier.eissn1573-0565en_NZ


Files in this item

This item appears in the following Collection(s)

Show simple item record