Biological Named Entity Recognition Using n-grams and Classification Methods

We propose a biological named entity recognition system which uses classification methods and a n-gram model to annotate terms in text. A novel method is presented to express lexical features in a pattern notation. Prefix and suffix characters are used instead of lists of potential terms or other external resources. Creating classification exemplars is conducted from text by using a word n-gram model. We evaluate our system based on the GENIA version 3.02 corpus which contains 2,000 paper abstracts. The system obtains an 0.705 F-score on exact match term performance. Biological concept markers are also assigned to each located term indicating its meaning. Our system retains simplicity and generalizability.