Evaluation of Sampling Methods for Learning from Imbalanced Data

Garima Goel, Liam Maguire, Yuhua Li, Sean McLoone

Research output: Chapter in Book/Report/Conference proceedingChapter

4 Citations (Scopus)

Abstract

The problem of learning from imbalanced data is of critical importance in a large number of application domains and can be a bottleneck in the performance of various conventional learning methods that assume the data distribution to be balanced. The class imbalance problem corresponds to dealing with the situation where one class massively outnumbers the other. The imbalance between majority and minority would lead machine learning to be biased and produce unreliable outcomes if the imbalanced data is used directly. There has been increasing interest in this research area and a number of algorithms have been developed. However, independent evaluation of the algorithms is limited. This paper aims at evaluating the performance of five representative data sampling methods namely SMOTE, ADASYN, BorderlineSMOTE, SMOTETomek and RUSBoost that deal with class imbalance problems. A comparative study is conducted and the performance of each method is critically analysed in terms of assessment metrics.
LanguageEnglish
Title of host publicationLecture Notes in Computer Science: Intelligent Computing Theories
Pages392-401
Volume7995
DOIs
Publication statusPublished - 2013

Fingerprint

Sampling
Learning systems

Cite this

Goel, G., Maguire, L., Li, Y., & McLoone, S. (2013). Evaluation of Sampling Methods for Learning from Imbalanced Data. In Lecture Notes in Computer Science: Intelligent Computing Theories (Vol. 7995, pp. 392-401) https://doi.org/10.1007/978-3-642-39479-9_47
Goel, Garima ; Maguire, Liam ; Li, Yuhua ; McLoone, Sean. / Evaluation of Sampling Methods for Learning from Imbalanced Data. Lecture Notes in Computer Science: Intelligent Computing Theories. Vol. 7995 2013. pp. 392-401
@inbook{b0ba6cd04a854defaef0e7754cae04fb,
title = "Evaluation of Sampling Methods for Learning from Imbalanced Data",
abstract = "The problem of learning from imbalanced data is of critical importance in a large number of application domains and can be a bottleneck in the performance of various conventional learning methods that assume the data distribution to be balanced. The class imbalance problem corresponds to dealing with the situation where one class massively outnumbers the other. The imbalance between majority and minority would lead machine learning to be biased and produce unreliable outcomes if the imbalanced data is used directly. There has been increasing interest in this research area and a number of algorithms have been developed. However, independent evaluation of the algorithms is limited. This paper aims at evaluating the performance of five representative data sampling methods namely SMOTE, ADASYN, BorderlineSMOTE, SMOTETomek and RUSBoost that deal with class imbalance problems. A comparative study is conducted and the performance of each method is critically analysed in terms of assessment metrics.",
author = "Garima Goel and Liam Maguire and Yuhua Li and Sean McLoone",
year = "2013",
doi = "10.1007/978-3-642-39479-9_47",
language = "English",
isbn = "978-3-642-39478-2",
volume = "7995",
pages = "392--401",
booktitle = "Lecture Notes in Computer Science: Intelligent Computing Theories",

}

Goel, G, Maguire, L, Li, Y & McLoone, S 2013, Evaluation of Sampling Methods for Learning from Imbalanced Data. in Lecture Notes in Computer Science: Intelligent Computing Theories. vol. 7995, pp. 392-401. https://doi.org/10.1007/978-3-642-39479-9_47

Evaluation of Sampling Methods for Learning from Imbalanced Data. / Goel, Garima; Maguire, Liam; Li, Yuhua; McLoone, Sean.

Lecture Notes in Computer Science: Intelligent Computing Theories. Vol. 7995 2013. p. 392-401.

Research output: Chapter in Book/Report/Conference proceedingChapter

TY - CHAP

T1 - Evaluation of Sampling Methods for Learning from Imbalanced Data

AU - Goel, Garima

AU - Maguire, Liam

AU - Li, Yuhua

AU - McLoone, Sean

PY - 2013

Y1 - 2013

N2 - The problem of learning from imbalanced data is of critical importance in a large number of application domains and can be a bottleneck in the performance of various conventional learning methods that assume the data distribution to be balanced. The class imbalance problem corresponds to dealing with the situation where one class massively outnumbers the other. The imbalance between majority and minority would lead machine learning to be biased and produce unreliable outcomes if the imbalanced data is used directly. There has been increasing interest in this research area and a number of algorithms have been developed. However, independent evaluation of the algorithms is limited. This paper aims at evaluating the performance of five representative data sampling methods namely SMOTE, ADASYN, BorderlineSMOTE, SMOTETomek and RUSBoost that deal with class imbalance problems. A comparative study is conducted and the performance of each method is critically analysed in terms of assessment metrics.

AB - The problem of learning from imbalanced data is of critical importance in a large number of application domains and can be a bottleneck in the performance of various conventional learning methods that assume the data distribution to be balanced. The class imbalance problem corresponds to dealing with the situation where one class massively outnumbers the other. The imbalance between majority and minority would lead machine learning to be biased and produce unreliable outcomes if the imbalanced data is used directly. There has been increasing interest in this research area and a number of algorithms have been developed. However, independent evaluation of the algorithms is limited. This paper aims at evaluating the performance of five representative data sampling methods namely SMOTE, ADASYN, BorderlineSMOTE, SMOTETomek and RUSBoost that deal with class imbalance problems. A comparative study is conducted and the performance of each method is critically analysed in terms of assessment metrics.

U2 - 10.1007/978-3-642-39479-9_47

DO - 10.1007/978-3-642-39479-9_47

M3 - Chapter

SN - 978-3-642-39478-2

VL - 7995

SP - 392

EP - 401

BT - Lecture Notes in Computer Science: Intelligent Computing Theories

ER -

Goel G, Maguire L, Li Y, McLoone S. Evaluation of Sampling Methods for Learning from Imbalanced Data. In Lecture Notes in Computer Science: Intelligent Computing Theories. Vol. 7995. 2013. p. 392-401 https://doi.org/10.1007/978-3-642-39479-9_47