Exemplar-Based Word Sense Disambiguation: Some Recent Improvements

TL;DR

This paper presents recent enhancements to exemplar-based word sense disambiguation, notably using larger k values and cross-validation, achieving accuracy comparable to the best existing methods on a large sense-tagged corpus.

Contribution

The paper introduces improvements to exemplar-based disambiguation by optimizing k and using cross-validation, resulting in higher accuracy on a large corpus.

Findings

Achieved higher disambiguation accuracy with larger k values.

Automatic k selection via cross-validation improves performance.

Comparable accuracy to the Naive-Bayes algorithm on the same dataset.

Abstract

In this paper, we report recent improvements to the exemplar-based learning approach for word sense disambiguation that have achieved higher disambiguation accuracy. By using a larger value of $k$, the number of nearest neighbors to use for determining the class of a test example, and through 10-fold cross validation to automatically determine the best $k$, we have obtained improved disambiguation accuracy on a large sense-tagged corpus first used in \cite{ng96}. The accuracy achieved by our improved exemplar-based classifier is comparable to the accuracy on the same data set obtained by the Naive-Bayes algorithm, which was reported in \cite{mooney96} to have the highest disambiguation accuracy among seven state-of-the-art machine learning algorithms.