search for




 

Confusion plot for the confusion matrix
Journal of the Korean Data & Information Science Society 2021;32:427-37
Published online March 31, 2021;  https://doi.org/10.7465/jkdi.2021.32.2.427
© 2021 Korean Data and Information Science Society.

Chong Sun Hong1

1Department of Statistics, Sungkyunkwan University
Correspondence to: 1Professor, Department of Statistics, Sungkyunkwan University, 25-2, Sungkyunkwan-Ro, Jongno-Gu, Seoul 03063, Korea. E-mail: cshong@skku.edu
Received December 8, 2020; Revised January 28, 2021; Accepted February 8, 2021.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
There is a well known 2 by 2 confusion matrix which is widely used in biostatistics or credit assessment. Using these, not only the true positive rate (TPR) and true negative rate (TNR) but also the positive predictive value (PPV) and negative predictive value (NPV) statistics are obtained. In this study, we propose the confusion plot which is a graphical method can geometrically describe various statistics defined from the confu- sion matrix. The confusion plot consists of six right-angled triangles using equations expressed as the sum of rows and columns of the confusion matrix. In particular, it is found that TPR, TNR, PPV and NPV can be described as acute angles of right-angled triangles in the confusion plot. This confusion plot can evaluate the classification model and its performance similarly to the well known ROC curve. Therefore, the confusion plot could be useful in evaluating the classi cation models similar to the ROC curve.
Keywords : Confusion matrix, NPV, PPV, TNR, TPR.