Open Access Article SciPap-886
On Reporting Performance of Binary Classifiers
by Pavel Škrabánek 1,* and Petr Doležel 2

1 Faculty of Electrical Engineering and Informatics, Department of Process Control, University of Pardubice, Studenská 95, Pardubice 532 10, Czechia

2 Faculty of Electrical Engineering and Informatics, Department of Process Control, University of Pardubice, Studenská 95, Pardubice 532 10, Czechia

* Authors to whom correspondence should be addressed.

Abstract: In this contribution, the question of reporting performance of binary classifiers is opened in context of the so called class imbalance problem. The class imbalance problem arises when a dataset with a highly imbalanced class distribution is used within the training or evaluation process. In such cases, only measures, which are not biased by distribution of classes in datasets, should be used; however, they cannot be chosen arbitrarily. They should be selected so that their outcomes provide desired information; and simultaneously, they should allow a full comparison of just evaluated classifier performance along, with performances of other solutions. As is shown in this article, the dilemma with reporting performance of binary classifiers can be solved using so called class balanced measures. The class balanced measures are generally applicable means, appropriate for reporting performance of binary classifiers on balanced as well as on imbalanced datasets. On the basis of the presented pieces of information, a suggestion for a generally applicable, fully-valued, reporting of binary classifiers performance is given.

Keywords: Machine Learning, Binary Classification, Class Imbalance Problem, Performance Measures, Reporting Of Results

JEL classification:   C45 - Neural Networks and Related Topics,   C83 - Survey Methods • Sampling Methods

SciPap 2017, 25(3), 886

Received: 16 December 2016 / Accepted: 23 October 2017 / Published: 5 December 2017