Comparative Studies on Some Morphological Analysis and Generation Techniques for Myanmar Language: A Review

Contenu principal de l'article

Kaung Myat Thu, H. Mamata Devi, Th. Rupachandra


Morphological Analysis And Generation (MAG) Play An Important Role In Natural Language Processing, Especially For Morphologically Rich Languages. It Is The Very First Step Toward Every NLP Task Such As Lemmatization, POS Tagging, Spell Checking, Grammar Checking, Machine Translation, Text Summarization, Information Extraction. MAG Deals With The Study Of Word Formation And Grammatical Structure Inside A Word. Every MAG Task Is Composed Of Three Main Parts: Morpheme Lexicon, Set Of Morphotactic Rules Or Orthographic Rules, And Decision Algorithms. In This Paper, We Have Reviewed Some Different Approaches Which Are Popular And Had Been Taken By Many Researchers. We Found That Corpus-Based Machine Learning Approach (SVM, NN, CRF, MDL, ...), Paradigm Based Approach, Two-Level Technique, Finite State Automata (FSA) Based Techniques, Finite State Transducers (FST) Based Techniques, Suffix Stripping, DAWG (Directed Acrylic Word Graph) Are Popular Successful Methods Reported In The Literature. Few Or Lack Of Researches And Developments In Morphological Analysis And Generation For The Myanmar Language Has Made This Study To Make A Review On The Literature Of Other Similar Languages.


Renseignements sur l'article