Feature Ranking Methods Used for Selection of Prototypes
Prototype selection, as a preprocessing step in machine learning, is effective in decreasing the computational cost of classification task by reducing the number of retained instances. This goal is obtained by shrinking the level of noise and rejecting the irrelevant data. Prototypes may be also used to understand the data through improving comprehensibility of results. In the paper we discus an approach for instance selection based on techniques known from feature selection pointing out the dualism between feature and instance selection. Finally some experiments are shown which uses feature ranking methods for instance selection.
KeywordsFeature Selection Mutual Information Feature Ranking Instance Selection Ranking Criterion
Unable to display preview. Download preview PDF.
- 2.Duch, W., Grudziński, K.: Prototype based rules - new way to understand the data. In: IEEE International Joint Conference on Neural Networks, pp. 1858–1863. IEEE Press, Washington D.C. (2001)Google Scholar
- 6.Aha, D., Kibler, D., Albert, M.K.: Instance-based learning algorithms. Machine Learning 6, 37–66 (1991)Google Scholar
- 7.Salvador, G., Joaquin, D., Cano, J.R., Herrera, F.: Prototype selection for nearest neighbor classification: Taxonomy and empirical study. IEEE Transactions on Pattern Analysis and Machine Intelligence 34(3), 417–435 (2010)Google Scholar
- 10.Duch, W., Wieczorek, T., Biesiada, J., Blachnik, M.: Comparision of feature ranking methods based on information entropy. In: Proc. of International Joint Conference on Neural Networks, pp. 1415–1420. IEEE Press, Budapest (2004)Google Scholar
- 12.Asuncion, A., Newman, D.:UCI machine learning repository (2007), http://www.ics.uci.edu/~mlearn/MLRepository.html
- 13.Duch, W., Maszczyk, T., Jankowski, N.: Make it cheap: learning with o(nd) complexity. In: 2012 IEEE World Congress on Computational Intelligence, Brisbane, Australia, pp. 132–135 (2012)Google Scholar
- 14.Bhattacharya, B.K., Poulsen, R.S., Toussaint, G.T.: Application of proximity graphs to editing nearest neighbor decision rule. In: Proc. Int. Symposium on Information Theory, Santa Monica, CA, pp. 1–25 (1981)Google Scholar