Comparison of Machine Learning Methods for Tamil Morphological Analyzer

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F22%3AFSDC288U" target="_blank" >RIV/00216208:11320/22:FSDC288U - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1007/978-981-16-2422-3_31" target="_blank" >https://doi.org/10.1007/978-981-16-2422-3_31</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-981-16-2422-3_31" target="_blank" >10.1007/978-981-16-2422-3_31</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Comparison of Machine Learning Methods for Tamil Morphological Analyzer

Popis výsledku v původním jazyce

Morphological Analysis is the study of word formation which explains how a word is evolved from smaller pieces called root word. Morphological analysis is an important task in natural language processing applications, namely, POS Tagging, Named Entity Recognition, Sentiment Analysis, and Information Extraction. The heart of the morphological analysis process is to find out the root words from the given documents that is exactly matched with the corpus list. There are many research works that have been done in this area of research however not much contribution has been made in domain specific in the area of domain specific analysis regional languages. Morphological analysis for regional languages is complex and demands extensive analysis of natural language rules and syntax pertaining to specific regional language of focus. In order to improve the quality of natural language processing, generally research works are restricted to domain specific analysis. Morphological analysis in Tamil language documents is quite complex and valuable for Tamil NLP process. Our work focuses on a comparative study of three different approaches in performing morphological analysis on the regional language called Tamil. The scope our work is restricted to Gynecology domain text in represented in Tamil language. The analysis of morphological process is done in three different machine learning methods for the Gynecological documents. The performance analysis is carried out on the three morphological analysis models, namely, Rules-based lemmatizer (IndicNLP), Paradigm-based Tamil Morphological Analyzer (Tacola), and N-gram-based lemmatizer (UDPipe), and our experimental results proved that paradigm-based finite state model gives optimal results (0.96).

Název v anglickém jazyce

Comparison of Machine Learning Methods for Tamil Morphological Analyzer

Popis výsledku anglicky

Morphological Analysis is the study of word formation which explains how a word is evolved from smaller pieces called root word. Morphological analysis is an important task in natural language processing applications, namely, POS Tagging, Named Entity Recognition, Sentiment Analysis, and Information Extraction. The heart of the morphological analysis process is to find out the root words from the given documents that is exactly matched with the corpus list. There are many research works that have been done in this area of research however not much contribution has been made in domain specific in the area of domain specific analysis regional languages. Morphological analysis for regional languages is complex and demands extensive analysis of natural language rules and syntax pertaining to specific regional language of focus. In order to improve the quality of natural language processing, generally research works are restricted to domain specific analysis. Morphological analysis in Tamil language documents is quite complex and valuable for Tamil NLP process. Our work focuses on a comparative study of three different approaches in performing morphological analysis on the regional language called Tamil. The scope our work is restricted to Gynecology domain text in represented in Tamil language. The analysis of morphological process is done in three different machine learning methods for the Gynecological documents. The performance analysis is carried out on the three morphological analysis models, namely, Rules-based lemmatizer (IndicNLP), Paradigm-based Tamil Morphological Analyzer (Tacola), and N-gram-based lemmatizer (UDPipe), and our experimental results proved that paradigm-based finite state model gives optimal results (0.96).

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

—

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Intelligent Sustainable Systems

ISBN

978-981-16-2422-3

ISSN

—

e-ISSN

—

Počet stran výsledku

15

Strana od-do

385-399

Název nakladatele

Springer

Místo vydání

—

Místo konání akce

Singapore

Datum konání akce

1. 1. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—