Weighting of Attributes in an Embedded Rough Approach
In an embedded approach to feature selection and reduction, a mechanism determining their choice constitutes a part of an inductive learning algorithm, as happens for example in construction of decision trees, artificial neural networks with pruning, or rough sets with activated relative reducts. The paper presents the embedded solution based on assumed weights for reducts and measures defined for conditional attributes, where weighting of these attributes was used in their backward elimination for rule classifiers induced in Dominance-Based Rough Set Approach. The methodology is illustrated with a binary classification case of authorship attribution.
Keywordsfeature selection reduction embedded approach DRSA reducts weighting authorship attribution stylometry
Unable to display preview. Download preview PDF.
- 1.Craig, H.: Stylistic analysis and authorship studies. In: Schreibman, S., Siemens, R., Unsworth, J. (eds.) A Companion to Digital Humanities. Blackwell Publishing Ltd., Oxford (2004)Google Scholar
- 2.Greco, S., Matarazzo, B., Słowiński, R.: Dominance-based rough set approach as a proper way of handling graduality in rough set theory. In: Peters, J.F., Skowron, A., Marek, V.W., Orłowska, E., Słowiński, R., Ziarko, W.P. (eds.) Transactions on Rough Sets VII. LNCS, vol. 4400, pp. 36–52. Springer, Heidelberg (2007)CrossRefGoogle Scholar
- 3.Jensen, R., Shen, Q.: Computational Intelligence and feature selection: Rough and fuzzy approaches. Wiley-IEEE Press (2008)Google Scholar