Relative Reduct-Based Estimation of Relevance for Stylometric Features
In rough set theory characteristic features, which describe classified objects, correspond to conditional attributes. A relative reduct is such an irreducible subset of attributes that preserves the quality of approximation of a complete decision table. For a decision table a single reduct or many reducts may exist. In typical processing one reduct is selected for the subsequent generation of decision rules, while others can be discarded. Yet when the set of reducts is analysed as a whole, observations and conclusions drawn can be used to evaluate relevance of attributes, which in turn can be employed in reduction of features not only for rule-based but also connectionist classifiers. The paper describes the steps of such procedure applied in the domain of stylometric processing of literary texts.
KeywordsDRSA Relative Reduct Characteristic Feature Relevance Measure Decision Algorithm ANN Stylometry
Unable to display preview. Download preview PDF.
- 1.Burrows, J.: Textual analysis. In: Schreibman, S., Siemens, R., Unsworth, J. (eds.) A Companion to Digital Humanities, ch. 23. Blackwell, Oxford (2004)Google Scholar
- 2.Craig, H.: Stylistic analysis and authorship studies. In: Schreibman, S., Siemens, R., Unsworth, J. (eds.) A Companion to Digital Humanities. Blackwell, Oxford (2004)Google Scholar
- 3.Fiesler, E., Beale, R.: Handbook of neural computation. Oxford University Press (1997)Google Scholar
- 4.Greco, S., Matarazzo, B., Sowiński, R.: The use of rough sets and fuzzy sets in Multi Criteria Decision Making. In: Gal, T., Hanne, T., Stewart, T. (eds.) Advances in Multiple Criteria Decision Making, ch. 14, pp. 14.1–14.59. Kluwer Academic Publishers, Dordrecht (1999)Google Scholar
- 5.Greco, S., Matarazzo, B., Sowiński, R.: Dominance-based rough set approach as a proper way of handling graduality in rough set theory. Transactions on Rough Sets 7, 36–52 (2007)Google Scholar