A review of word-sense disambiguation methods and algorithms: Introduction

Abstract

The word-sense disambiguation task is a classification task, where the goal is to predict the meaning of words and phrases with the help of surrounding text. The purpose of this short review is to acquaint the reader with the general directions of word-sense disambiguation methods and algorithms. These approaches include the following groups of methods: neural network, machine learning meta-algorithms (AdaBoost), lexical chain computation, methods based on Bayes' theorem, context clustering and words clustering algorithms. The experimental comparison of different algorithms concludes this review. This paper is licensed under the CC Attribution license.

Keywords

• word-sense disambiguation
• neural network
• boosting
• lexical chain
• naive bayes classifier
• bayesian network
• selectional preferences
• word-sense discrimination