A Customized Adaptation of Traditional Lesk Method for Sense Disambiguation of Punjabi Words from Medical Domain
Jaskiran Kaur, Amardeep S. G. Singh · 2015
We report in this paper a way of doing Word Sense Disambiguation (WSD) that has its origin in multilingual MT. Various Machine Learning approaches have been demonstrated to produce relatively successful Word Sense Disambiguation systems. For English and other European languages, a lot of research has been done. Indian languages are still intact in this area. A WordNet for Indian languages, IndoWordNet was developed by IIT Bombay which is the only source of information when it comes to Indian languages. We have proposed a algorithm for disambiguating polysemous words in Punjabi words particularly belonging to medical domain and compared the results with the classical Lesk algorithm.