SweNam-A Swedish Named Entity recognizer Its construction, training and evaluation
Hercules Dalianis, Erik Åström · 2001
In this paper we describe the development, training and evaluation of a Swedish Named Entity (NE) tagger called SweNam. NE tagging or recognition is the technique where one extracts words describing Persons, Locations, Organizations and Time from a text. We have used a combination of Machine Learning (ML) techniques and matching rules to construct our NE tagger. The training corpus consists of 108 000 Swedish news articles downloaded from Internet during 2000-2001 and we have used a number of ready NE lexicons to bootstrap our system.