Generic performance measure for multiclass-classifiers

Kautz T, Eskofier B, Pasluosta CF (2017)


Publication Language: English

Publication Type: Journal article, Original article

Publication year: 2017

Journal

Book Volume: 68

Pages Range: 111 - 125

Journal Issue: 8

DOI: 10.1016/j.patcog.2017.03.008

Abstract

The evaluation of classification performance is crucial for algorithm and model selection. However, a performance measure for multiclass classification problems (i.e., more than two classes) has not yet been fully adopted in the pattern recognition and machine learning community. In this work, we introduce the multiclass performance score (MPS), a generic performance measure for multiclass problems. The MPS was designed to evaluate any multiclass classification algorithm for any arbitrary testing condition. This measure handles the case of unknown misclassification costs and imbalanced data, and provides confidence indicators of the performance estimation. We evaluated the MPS using real and synthetic data, and compared it against other frequently used performance measures. The results suggest that the proposed MPS allows capturing the performance of a classification with minimum influence from the training and testing conditions. This is demonstrated by its robustness towards imbalanced data and its sensitivity towards class separation in feature space.

Authors with CRIS profile

Additional Organisation(s)

How to cite

APA:

Kautz, T., Eskofier, B., & Pasluosta, C.F. (2017). Generic performance measure for multiclass-classifiers. Pattern Recognition, 68(8), 111 - 125. https://dx.doi.org/10.1016/j.patcog.2017.03.008

MLA:

Kautz, Thomas, Björn Eskofier, and Cristian Federico Pasluosta. "Generic performance measure for multiclass-classifiers." Pattern Recognition 68.8 (2017): 111 - 125.

BibTeX: Download