A novel oversampling technique based on the manifold distance for class imbalance learning

dc.cclicenceN/Aen
dc.contributor.authorGuo, Yinan
dc.contributor.authorJiao, Botao
dc.contributor.authorYang, Lingkai
dc.contributor.authorCheng, Jian
dc.contributor.authorYang, Shengxiang
dc.contributor.authorTang, Fengzhen
dc.date.acceptance2020-08-06
dc.date.accessioned2020-09-08T10:08:09Z
dc.date.available2020-09-08T10:08:09Z
dc.date.issued2020-11-25
dc.descriptionThe file attached to this record is the author's final peer reviewed version.en
dc.description.abstractOversampling is a popular problem-solver for class imbalance learning by generating more minority samples to balance the dataset size of different classes. However, resampling in original space is ineffective for the imbalance datasets with class overlapping or small disjunction. Based on this, a novel oversampling technique based on manifold distance is proposed, in which a new minority sample is produced in terms of the distances among neighbors in manifold space, rather than the Euclidean distance among them. After mapping the original data to its manifold structure, the overlapped majority and minority samples will lie in areas easily being partitioned. In addition, the new samples are generated based on the neighbors locating nearby in manifold space, avoiding the adverse effect of the disjoint minority classes. Following that, an adaptive adjustment method is presented to determine the number of the newly generated minority samples according to the distribution density of the matched-pair data. The experimental results on 48 imbalanced datasets indicate that the proposed oversampling technique has the better classification accuracy.en
dc.funderOther external funder (please detail below)en
dc.funder.otherNational Natural Science Foundation of Chinaen
dc.identifier.citationGuo, Y., Jiao, B., Cheng, J., Yang, L., Yang, S., and Tang, F. (2020) A novel oversampling technique based on the manifold distance for class imbalance learning. International Journal of Bio-Inspired Computation, 18 (3), pp. 131-142en
dc.identifier.doihttps://doi.org/10.1504/IJBIC.2021.119197
dc.identifier.issn1758-0366
dc.identifier.urihttps://dora.dmu.ac.uk/handle/2086/20156
dc.language.isoen_USen
dc.peerreviewedYesen
dc.projectid61973305en
dc.projectid61573361en
dc.projectid61803369en
dc.publisherInderscience Publishers Ltden
dc.researchinstituteInstitute of Artificial Intelligence (IAI)en
dc.subjectClass imbalance learningen
dc.subjectoversamplingen
dc.subjectmanifold learningen
dc.subjectoverlappingen
dc.subjectsmall disjunctionen
dc.titleA novel oversampling technique based on the manifold distance for class imbalance learningen
dc.typeArticleen

Files

Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
IJBIC20.pdf
Size:
1.35 MB
Format:
Adobe Portable Document Format
Description:
Main article
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
4.2 KB
Format:
Item-specific license agreed upon to submission
Description: