Disambiguating Multi-Word Expressions (MWEs) is often a critical task in NLP applications.
Reduplications are an important subclass of MWEs and they are a high-frequency occurrence
compared to other kinds of MWEs in Hindi. There are some linguistic challenges in classification
and identification of Reduplicated Multiword Expressions (RMWEs) in Hindi. The aim of this
paper is to demonstrate linguistic issues pertaining to the distribution of RMWEs, their
formalization aspects using a CRF based CRF++ tool and testing and evaluation of the trained
system. As per our knowledge, there is no available guideline for annotation of MWEs in Hindi.
Therefore, we are presenting the first detailed guidelines for annotation of MWEs in Hindi and it
can be applicable in other Indian Languages as well.
Added on June 8, 2016
Contributed by : Atul
Product Type : Research Paper
License Type : Freeware
System Requirement :
Author : Renu Singh, Atul Kr. Ojha and Girish Nath Jha