Morpheme Boundary Identification Using Letter Successor Variety

Morpheme boundary identification is one of the prominent problems in morphological analysis of NLP applications. This study proposes an alternative technique to effectively identify the boundary of each morpheme from a compound word. This is useful in wide range of NLP applications from stemming, word assistance to document categorizers. Malayalam, a major South Indian language has the linguistic capability to have high number of morphemes per word. The present study sets out to discover the effectiveness of Letter Successor Variety techniques could be useful for identifying the morpheme boundary in Malayalam. Letter Successor Variety is based on statistical co-occurrence measures and contextually similar words.