Assessments Metrics for Multi-class Imbalance Learning: A Preliminary Study
In this paper we study some of the most common global measures employed to measure the classifier performance on the multi-class imbalanced problems. The aim of this work consists of showing the relationship between global classifier performance (measure by global measures) and partial classifier performance, i.e., to determine if the results of global metrics match with the improved classifier performance over the minority classes. We have used five strategies to deal with the class imbalance problem over five real multi-class datasets on neural networks context.
KeywordsMulti-class imbalance global measures accuracy by class
- 1.Tsoumakas, G., Katakis, I.: Multi-label classification: An overview. Int. J. Data Warehousing and Mining, 1–13 (2007)Google Scholar
- 3.Wang, S., Yao, X.: Multi-class imbalance problems: Analysis and potential solutions. IEEE Transactions on IEEE Transactions on Systems, Man and Cybernetics, Part B: Cybernetics (99), 1–12 (2012)Google Scholar
- 7.García, V., Mollineda, R.A., Sánchez, J.S.: Theoretical analysis of a performance measure for imbalanced data. In: ICPR, pp. 617–620 (2010)Google Scholar
- 11.A. Asuncion, D.N.: UCI machine learning repository (2007)Google Scholar
- 13.Weiss, G.M., Provost, F.J.: Learning when training data are costly: The effect of class distribution on tree induction. J. Artif. Intell. Res. (JAIR) 19, 315–354 (2003)Google Scholar