Product Named Entity Recognition Using Conditional Random Fields

Abstract

For Product Named Entity Recognition, Conditional Random Fields model is used in this paper. By introducing the domain ontology features to the CRFs model can improve the performance of Product Named Entity Recognition, and experiments were made to compare the two kinds of feature templates. Empirical results show that the PNER based on CRFs model with ontology features can achieve good performance. However, the Imperfect of Domain Ontology leads to the product named entity recognition may be not better than the research of other previous scholars using other methods for traditional named entity recognition.