Subject and object identification in Malayalam text
Saranya D. Krishnan, R R Rajeev, Mary Priya Sebastian, Elizabeth Sherly · 2012
Subject and object identification denotes the process of identifying the syntactic subject and object in a sentence while tagging. Identification of subject/object turned out as very effective for many natural language applications such as machine translation, anaphora resolution, relationship extraction etc. For languages with rigid word order, the subject, object and verb in a sentence can be identified by their position. But for languages like Malayalam with relatively free word order the subject/object identification task is more complex and is significant. This paper presents a method for subject and object identification in Malayalam text using statistical tagging approach using HMM and Viterbi algorithm. In this approach, the tagset for part of speech tagging is modified to include tags that distinguish between the subject and object in a sentence.