JU_CSE_GREC10: Named Entity Generation at GREC 2010

This paper presents the experiments carried out at Jadavpur University as part of the participation in the GREC Named Entity Generation Challenge 2010. The Baseline system is based on the SEMCAT, SYNCAT and SYNFUNC features of REF and REG08-TYPE and CASE features of REFEX elements. The discourse level system is based on the additional positional features: paragraph number, sentence number, word position in the sentence and mention number of a particular named entity in the document. The inclusion of discourse level features has improved the performance of the system. 1 Baseline System The baseline system is based on the following linguistic features of REF elements: SEMCAT (Semantic Category), SYNCAT (Syntactic Category) and SYNFUNC (Syntactic Function) (Anja Belz, 2010) and the following linguistic features of REFEX elements: REG08-TYPE (Entity type) and CASE (Case marker). The baseline system has been separately trained on the training set data for the three domains: chefs, composer...

• Type: Journal
• Year: 2010
• Where: INLG
• Authors: Amitava Das, Tanik Saikh, Tapabrata Mondal, Sivaji Bandyopadhyay
• Keywords: INLG 2010, Natural Language Processing, Probable Refex Element, Ref Element, Refex Element