Morpheme Extraction in Tamil using Finite State Machines (FIRE-2013 - Morpheme Extraction Task)
Sobha Lalitha Devi, K Marimuthu, Vijay Sundar, Ram Ratan, K Amudha · 2013
In this work we present an efficient morphological analysis system for Tamil where we extract the individual morphemes using Finite State Automaton (FSA). The aim is to perform morpheme analysis by modeling the regular inflectional pattern exhibited by Tamil as an FSA. We handle the compound and agglutinated words using CRF based word boundary identifier. On linguistic evaluation the system achieved an encouraging accuracy of 86.17%