An Affix Based Word Classification Method of Assamese Text

Bhairab Sarma, Bipul Shyam Purkayastha


Classification of word is an important activity in Natural Language Processing (NLP) analysis. Word classification as we mean in linguistic is not same as in natural language processing. In NLP, the main objective is Part-of-Speech tagging (POST) which if essential for machine translation and language interpretation. However, in linguistic, words are classified as their applications and representation of meaning in the context of real world. Retrieving contextual meaning in language processing is a very challenging job. Because of sense disambiguation, representation ambiguity and words with multiple meaning, the task POST become very difficult.Assamese is a highly inflected and morphologically rich Indian language. In this study, we attempt to classify words based on its morphological structure. We present a method of classification of Assamese word based on its inflectional features. The classes we have used here may not be complement with POS classification. However it could be method of word clustering during POS with application of other smoothing algorithm like HMM, EM etc. We believe that this method can further be implementing into any other inflectional Indian language processing.

Keywords: Affixes, Contextual meaning, NLP, POST, WSD

Full Text:




  • There are currently no refbacks.

Copyright (c) 2016 International Journal of Advanced Research in Computer Science