Multi-objective Optimisation-Based Feature Selection for Multi-label Classification
In this short note we introduce multi-objective optimisation for feature subset selection in multi-label classification. We aim at optimise multiple multi-label loss functions simultaneously, using label powerset, binary relevance, classifier chains and calibrated label ranking as the multi-label learning methods, and decision trees and SVMs as base learners. Experiments on multi-label benchmark datasets show that the feature subset obtained through MOO performs reasonably better than the systems that make use of exhaustive feature sets.
KeywordsMulti-label classification Multi-objective optimisation Feature subset selection
- 1.Chang, C.C., Lin, C.J.: LibSVM: a library for support vector machines. ACM Trans. Intell. Syst. Technol. 2(3) (2011). Article No. 27Google Scholar
- 2.Deb, K.: Multi-objective Optimization Using Evolutionary Algorithms. Wiley, Chichester (2001)Google Scholar
- 3.Khan, M.A., Ekbal, A., Loza Mencía, E., Fürnkranz, J.: Multi-objective optimisation-based feature selection for multi-label classification. Technical report TUD-KE-2017-01, Knowledge Engineeering Group, TU Darmstadt (2017)Google Scholar
- 4.Quinlan, J.R.: C4.5: Programs for Machine Learning. Morgan Kaufmann Publishers Inc., San Francisco (1993)Google Scholar
- 6.Tsoumakas, G., Katakis, I., Vlahavas, I.P.: Mining multi-label data. In: Data Mining and Knowledge Discovery Handbook, 2nd edn., pp. 667–685 (2010)Google Scholar