Machine Learning with Lexical Features: The Duluth Approach to Senseval-2

TL;DR

This paper presents the Duluth systems' entries in Senseval-2, utilizing standard machine learning algorithms with lexical features for word sense disambiguation in English and Spanish.

Contribution

Introduces portable, robust machine learning methods based on lexical features for word sense disambiguation, demonstrating their effectiveness across languages.

Findings

High portability and robustness of the methods

Effective disambiguation in English and Spanish

Foundation for more tailored approaches

Abstract

This paper describes the sixteen Duluth entries in the Senseval-2 comparative exercise among word sense disambiguation systems. There were eight pairs of Duluth systems entered in the Spanish and English lexical sample tasks. These are all based on standard machine learning algorithms that induce classifiers from sense-tagged training text where the context in which ambiguous words occur are represented by simple lexical features. These are highly portable, robust methods that can serve as a foundation for more tailored approaches.