Class imbalance and cost-sensitive decision trees: A unified survey based on a core similarity

Michael Siers, Zahid Islam

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

Class imbalance treatment methods and cost-sensitive classification algorithms are typically treated as two independent research areas. However, many of these techniques have properties in common. After providing a background to the two fields of research, this article identifies the fundamental mechanism which is common to both. Using this mechanism, a taxonomy is created which encompasses approaches to both class imbalance treatment and cost-sensitive classification. Through this survey, we aim to bridge the gap between the two fields such that lessons from one field may be applied to the other. Many data mining tasks are naturally both class imbalanced and cost-sensitive. This survey is useful for researchers and practitioners approaching these tasks as it provides a detailed overview of approaches in both fields. Many of the surveyed techniques are classifier independent. However, we chose to focus on techniques which were either decision tree-based or compatible with decision trees. This choice was based on the popularity and novelty of their application to both fields.
Original languageEnglish
Article number4
Pages (from-to)1- 31
Number of pages31
JournalACM Transactions on Knowledge Discovery from Data
Volume15
Issue number1
DOIs
Publication statusPublished - Dec 2020

Fingerprint

Dive into the research topics of 'Class imbalance and cost-sensitive decision trees: A unified survey based on a core similarity'. Together they form a unique fingerprint.

Cite this