سال انتشار: ۱۳۸۶

محل انتشار: پانزدهیمن کنفرانس مهندسی برق ایران

تعداد صفحات: ۵

نویسنده(ها):

Mandana Hamidi – Department of Computing and Information Technology , Azad University of Qazvin
Ali Borji – School of Cognitive Sciences, Institute for Studies in Theoretical Physics and Mathematics
Saeed Shiry Ghidary – Computer Engineering Department, Amirkabir University of Technology

چکیده:

Word Sense Disambiguation (WSD) aims to identify the correct sense of an ambiguous word in a sentence. In thiswork we investigate the performance of two state – of – the art approaches: k-NN and Naïve Bayes for the purpose of Persian word sense disambiguation. These methods have been evaluated on two highly frequent and ambiguous words from “Hamshahri” – by some means a standard corpus for Persian language. We performed experiments on both stemmed and non-stemmed version of the corpus. The results show the superiority of k-NN algorithm over Naïve Bayes in almost all case. Although the results demonstrate good performance, furtherinvestigation should be done , by trying other classification meghods and also other features used in the literature.