Automated Subcategorization of Named Entities

There has been much interest in the recent past concerning the possibilities for automated categorization of named entities. The research presented here describes a method for the subcategorization of location names. Subcategorization of locations is not a trivial task even for human subjects, who perform at accuracy levels of less than 58%. After experimenting with both Bayesian classifiers and decision tree learning algorithms, we have designed a system that achieves accuracy levels greater than 80%.

• Type: Conference
• Year: 2001
• Where: ACL
• Authors: Michael Fleischman
• Keywords: Accuracy Levels, ACL 2001, ACL 2007, Decision Tree Learning, Trivial Task