A hybrid approach to NER by MEMM and manual rules

This paper describes a framework for defining domain specific Feature Functions in a user friendly form to be used in a Maximum Entropy Markov Model (MEMM) for the Named Entity Recognition (NER) task. Our system called MERGE allows defining general Feature Function Templates, as well as Linguistic Rules incorporated into the classifier. The simple way of translating these rules into specific feature functions are shown. We show that MERGE can perform better from both purely machine learning based systems and purely-knowledge based approaches by some small expert interaction of rule-tuning. Categories and Subject Descriptors I.2.7 [Natural Language Processing]: Text analysis – Named Entity Recognition, Information Extraction. General Terms Named Entity, Information, Document Collection, Statistical Model, Empirical Model, Machine Learning, Maximum Entropy. Keywords Named Entity Recognition, Text Mining, Machine Learning, Information Extraction, Maximum Entropy Markov Model.

• Type: Conference
• Year: 2005
• Where: CIKM
• Authors: Moshe Fresko, Binyamin Rosenfeld, Ronen Feldman
• Keywords: CIKM 2005, Entropy Markov Model, Maximum Entropy Markov, Named Entity Recognition