Weighted Combination of Classifiers Based on Dempster-Shafer Theory and OWA Operators in Word Sense Disambiguation

In this paper we discuss a framework for weighted combination of classifiers in which each individual classifier uses a distinct representation of objects to be classified. This framework is essentially based on Dempster-Shafer theory of evidence (Dempster, 1967; Shafer, 1976) and OWA operators (Yager, 1988). It is of interest to see that this framework not only yields many commonly used decision rules without some strong assumptions made in the work by Kittler et al. (1998), but also provides other new decision rules. As an application, we apply the proposed framework of classifier combination to the problem of word sense disambiguation (shortly, WSD). To this end, we experimentally design a set of individual classifiers, each of which corresponds to a distinct representation type of context considered in the WSD literature, and then the discussed combination strategies are tested on the datasets for four polysemous words, namely interest, line, serve, and hard. The experiment conducted for these four polysemous words shows significantly better results in comparison with previous studies on the same datasets.The original publication is available at JAIST Press http://www.jaist.ac.jp/library/jaist-press/index.htmlIFSR 2005 : Proceedings of the First World Congress of the International Federation for Systems Research : The New Roles of Systems Sciences For a Knowledge-based Society : Nov. 14-17, 2120, Kobe, JapanSymposium 5, Session 3 : Data/Text Mining from Large Databases Data Minin