IIIT Hyderabad Publications |
|||||||||
|
RUSSE’2018: WORD SENSE INDUCTION AND DISAMBIGUATION METHOD BASED ON CONTEXT-BASED LISTSAuthors: Sreya Mittal,Pratibha Rani Conference: Dialogue-2018 (Dialogue-2018 2018) Date: 2018-05-30 Report no: IIIT/TR/2018/36 AbstractThis paper reports the participation of IIITHDSAC team in the shared task on word sense induction and disambiguation (WSID) for the Russian language in RUSSE’2018. The method adopted is semi-supervised and knowledge-free which does not use any knowledge resource like dictionary or Wiki. It only uses the sense tagged and untagged data provided by the task organizers as training data and builds the WSID model using the concept of context-based lists from words of training data converted into root form. Context-based lists enables to cluster words and senses based on contexts and hence, provides a way to use context of an unseen target word to find its sense even if it is absent in the training data. We have used the root form of training set words because the test set words were given in root form otherwise our method is generic and would work for normal form of words also. Full paper: pdf Centre for Data Engineering |
||||||||
Copyright © 2009 - IIIT Hyderabad. All Rights Reserved. |