Improved Named Entity Tagset for Punjabi Language

Evaluation of Punjabi Named Entity Recognition using Context Word Feature

“…As the annotation task progresses, more ambiguous and confusing cases are identified that do not fit the given tagset and annotation guidelines thus leads to inconsistency in the corpus. In [7], various tagset design issues and problematic cases were discussed and accordingly proposed additional tags to be used for NER task in Punjabi language. In order to develop the tagset, Extended Named Entity hierarchy provided in [15], CoNLL 2002 and 2003 tagset and 12 tags of NERSSEAL were referred.…”

“…During this work various ambiguous and problematic cases were realized. Various tagset design issues and author's recommendations about additional tags were presented and finally a tagset of 14 tags have been proposed in [7] which will be used in the current research work. Although Extended Named Entity hierarchy provides more than 100 tags but we have opted a limited tagset.…”