Automated Named Entity Recognition from Tamil Documents

Abstract

Named Entity Recognition (NER) is a subsequence of words in a document that seeks to detect and classify entities into pre-defined categories such as name of the person, organization and location respectively. The impact on NER is high because a lot of Information Extraction (IE) relations are associated using Named Entities (NEs). This paper presents a pioneering method for extraction of NEs for Tamil using Supervised Learning. This hybrid framework makes use of features that are extracted based on the speciality of the Tamil language NEs. The evaluation has been done by using 1028 number of documents which comprises the standard FIRE corpus and an F-measure of 83.54% has been achieved. A performance comparison with one of the state-of-the-art Tamil NE system has been done and the proposed methodology has achieved better accuracy.