Bengali Noun Morphological Analyzer

This paper describes a linguistics approach towards development of a Bengali Noun Morphological Analyzer implemented at first on the semi-manually created database of 87697 inflected words list tokens, i.e. Input2 for Linguistics Resource Creation comprising of Noun, Pronoun, Adjective roots with and without its suffixes. Then after the first implementation the developed Linguistic Resource knowledge is applied on an unknown Bengali corpus database containing 6157 tokens. At the initial stage of this research a linguistic analysis is done which leads to framing of the nominal suffix list which is later on used in nominal suffix extraction. This linguistic knowledge is implemented in developing the finite-state transducer grammar for Linguistic Resource which gives way to the development of Bengali Noun Morphological Analyzer. The final output obtained is around 44% accuracy. This accuracy can be always improved with time if we keep on increasing the nominal roots in the FST grammar file.