A Deeper Study on Features for Named Entity Recognition

Malarkodi C S, Sobha Lalitha Devi


Abstract
This paper deals with the various features used for the identification of named entities. The performance of the machine learning system heavily depends on the feature selection criteria. The intention to trace the essential features required for the development of named entity system across languages motivated us to conduct this study. The linguistic analysis was done to find out the part of speech patterns surrounding the context of named entities and from the observation linguistic oriented features are identified for both Indian and European languages. The Indian languages belongs to Dravidian language family such as Tamil, Telugu, Malayalam, Indo-Aryan language family such as Hindi, Punjabi, Bengali and Marathi, European languages such as English, Spanish, Dutch, German and Hungarian are used in this work. The machine learning technique CRFs was used for the system development. The experiments were conducted using the linguistic features and the results obtained for each languages are comparable with state-of-art systems.
Anthology ID:
2020.wildre-1.12
Volume:
Proceedings of the WILDRE5– 5th Workshop on Indian Language Data: Resources and Evaluation
Month:
May
Year:
2020
Address:
Marseille, France
Venues:
LREC | WILDRE | WS
SIG:
Publisher:
European Language Resources Association (ELRA)
Note:
Pages:
66–72
Language:
English
URL:
https://www.aclweb.org/anthology/2020.wildre-1.12
DOI:
Bib Export formats:
BibTeX MODS XML EndNote
PDF:
http://aclanthology.lst.uni-saarland.de/2020.wildre-1.12.pdf