Manipuri Morpheme Identification

The Morphemes of the Manipuri word are the real bottleneck for any of the Manipuri Natural Language Processing (NLP) works. It is one of the Indian Scheduled Language with less advancement so far in terms of NLP applications. This is because the nature of the language is highly agglutinative. Segmentation of a word and identifying the morphemes becomes necessary before proceeding for any of the Manipuri NLP application. A highly inflected word may sometimes consist of ten or more affixes. These affixes are the morphemes which change the semantic and grammatical structure. So the inflexion in a word plays an important role. Words are segmented to the syllables and are examined to extract a morpheme among the syllables. This work is implemented in the Manipuri words written with the Meitei Mayek (script). This is because the syllable formations are distinct comparing to the Manipuri written with Bengali script. The combination of 2-gram or bi-gram and the Standard Deviation technique are used for the identification of the morphemes. This system gives an output with the recall of 59.80%, the precision of 83.02% and the f-score of 69.52%.