A multilingual clause boundary detection approach for Assamese and Bishnupriya Manipuri

Clause boundary detection is an important step in various natural language processing tasks like dependency parsing, machine translation etc. A complex sentence in any language may contain more than one clause. The boundaries should be identified for capturing various dependency relations and for identifying the arguments of the verbal components. We present here, the task of identification and classification of clauses in Assamese and Bishnupriya Manipuri text. To the best of our knowledge, not much work has been done on clause boundary identification for these two languages, which makes this task more important. We have built a rule-based system using linguistic cues such as coordinating conjunct, subordinating conjunct etc. We additionally considered POS tags of neighbouring words as features into the existing conjunction-based approach, which can also be applied to the languages that belongs to the same language family and have certain similarities. Experimental results show that our approach achieves a satisfactory result.