Automatic identification of assamese and bodo multiword expressions

Multiword Expressions (MWEs) are sequence of words separated by space or delimiter which determines a unique meaning instead of words´ individual meanings. Our work concentrates on automatic identification of MWEs for two less computationally aware languages Assamese and Bodo spoken in the North Eastern part of India. Statistical measure and Language specific knowledge helps us to extract MWEs from raw corpus. Natural Language Processing tasks in Assamese and Bodo languages have started in recent years, and this is the first organised approach to exploit MWEs in both these languages. Linguistics aspects for analysing the results have been considered, and we have found the results quite satisfactory.