Publication Type : Journal Article
Publisher : Journal of Advanced Research in Dynamical Control Systems .
Source : Journal of Advanced Research in Dynamical & Control Systems, Volume 10, Number 8 (2018)
Campus : Kochi
School : School of Arts and Sciences
Department : English
Year : 2018
Abstract : Machine translation (MT) is an applied form of computational linguistics and Artificial intelligence (AI).In an MT context the realization of the exact sense expressed by each and every lexical element in an input sentence by a machine is very important. Sense disambiguation is one of the hardest tasks for a machine, because most of the lexical items of any natural language show polysemy and synonymy. The nature of polysemy of lexical items is one of the major causes for negative transfer (wrong transfer) in MT context. This study focuses on the polysemy nature of Simple English Prepositions (SEP) in an English to Hindi MT context. The transfer of any Simple English Preposition (SEP) into any other language by a machine is very difficult, because the Simple English prepositions (at, by, from, for, in, of, on, to and with)are highly ambiguous. These prepositions produce different linguistic senses with the same syntax. The various linguistic senses produced by simple English prepositions are transformed into different postpositions in Hindi. Finding the exact translation equivalences in Hindi for Simple English Prepositions (SEPs) in the MT context still remains a complicated task. This study proposes a new methodology for machine learning (ML)approach, which utilizes the linguistic features of the predicative part or the Verb Phrase (VP) of a sentence. This study takes advantage of Levin�s classification of English verbs and the researcher�s own noun classification for extracting the linguistic features of the main verb in the VP and the nouns in the predicative part of the sentences. A supervised machine learning approach (SVM) is used for the sense identification of the simple English prepositions �to� and �with� in an English to Hindi MT context.
Cite this Research Publication : J. D. Ratnam, Soman, K. P., B. Premjith, and Dr. M. G. Priya, “Transfer of Simple English Prepositions ‘to’ and ‘with’ Into Hindi Utilizing Linguistic Features of the Predicative Part of a Sentence with Machine Learning Approach in an English to Hindi MT Context”, Journal of Advanced Research in Dynamical & Control Systems, vol. 10, 2018.