On the performance of Matthews correlation coefficient (MCC) for imbalanced dataset

Qiuming Zhu

doi:10.1016/j.patrec.2020.03.030

On the performance of Matthews correlation coefficient (MCC) for imbalanced dataset

Qiuming Zhu

Computer Science

Research output: Contribution to journal › Article › peer-review

132 Scopus citations

Abstract

The Matthews Correlation Coefficient (MCC) is one of the popular measurements for classification accuracy. It has been generally regarded as a balanced measure which can be used even if the classes are of very different sizes. The study of this paper finds that this is not true. MCC deteriorates seriously when the dataset in classification are imbalanced. Experiment results and analysis show that MCC is not suitable for classification accuracy measurement on imbalanced datasets.

Original language	English (US)
Pages (from-to)	71-80
Number of pages	10
Journal	Pattern Recognition Letters
Volume	136
DOIs	https://doi.org/10.1016/j.patrec.2020.03.030
State	Published - Aug 2020

Keywords

Classification accuracy measurement
Imbalanced dataset
Matthews correlation coefficient
Performance evaluation

ASJC Scopus subject areas

Software
Signal Processing
Computer Vision and Pattern Recognition
Artificial Intelligence

Access to Document

10.1016/j.patrec.2020.03.030

Cite this

@article{bd914c87d4874db9b59f4264400339c7,

title = "On the performance of Matthews correlation coefficient (MCC) for imbalanced dataset",

abstract = "The Matthews Correlation Coefficient (MCC) is one of the popular measurements for classification accuracy. It has been generally regarded as a balanced measure which can be used even if the classes are of very different sizes. The study of this paper finds that this is not true. MCC deteriorates seriously when the dataset in classification are imbalanced. Experiment results and analysis show that MCC is not suitable for classification accuracy measurement on imbalanced datasets.",

keywords = "Classification accuracy measurement, Imbalanced dataset, Matthews correlation coefficient, Performance evaluation",

author = "Qiuming Zhu",

note = "Publisher Copyright: {\textcopyright} 2020 Elsevier B.V.",

year = "2020",

month = aug,

doi = "10.1016/j.patrec.2020.03.030",

language = "English (US)",

volume = "136",

pages = "71--80",

journal = "Pattern Recognition Letters",

issn = "0167-8655",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - On the performance of Matthews correlation coefficient (MCC) for imbalanced dataset

AU - Zhu, Qiuming

PY - 2020/8

Y1 - 2020/8

N2 - The Matthews Correlation Coefficient (MCC) is one of the popular measurements for classification accuracy. It has been generally regarded as a balanced measure which can be used even if the classes are of very different sizes. The study of this paper finds that this is not true. MCC deteriorates seriously when the dataset in classification are imbalanced. Experiment results and analysis show that MCC is not suitable for classification accuracy measurement on imbalanced datasets.

AB - The Matthews Correlation Coefficient (MCC) is one of the popular measurements for classification accuracy. It has been generally regarded as a balanced measure which can be used even if the classes are of very different sizes. The study of this paper finds that this is not true. MCC deteriorates seriously when the dataset in classification are imbalanced. Experiment results and analysis show that MCC is not suitable for classification accuracy measurement on imbalanced datasets.

KW - Classification accuracy measurement

KW - Imbalanced dataset

KW - Matthews correlation coefficient

KW - Performance evaluation

UR - http://www.scopus.com/inward/record.url?scp=85086413445&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85086413445&partnerID=8YFLogxK

U2 - 10.1016/j.patrec.2020.03.030

DO - 10.1016/j.patrec.2020.03.030

M3 - Article

AN - SCOPUS:85086413445

SN - 0167-8655

VL - 136

SP - 71

EP - 80

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

ER -

On the performance of Matthews correlation coefficient (MCC) for imbalanced dataset

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this