Abstract | ||
---|---|---|
A framework for product named entity recognition in Chinese was presented using Conditional Random Fields with multiple features in this paper. It differentiates from most of the previous approaches mainly as follows. Firstly, introducing the domain ontology features to the CRFs model can use its semantic information. Secondly, combining internal and external features to compound features can use more rich overlapping features. so that it can improve the performance of product named entity Recognition. Experimental results show that this approach can achieve an overall F-measure around 87.16%, which seems to achieve the current state-of-the-art performance. However, due to the imperfect of Domain Ontology and the complication of reviews texts, the recognition for product named entity may not be better than the research of the traditional named entity recognition. |
Year | DOI | Venue |
---|---|---|
2012 | 10.1109/CSCWD.2012.6221863 | CSCWD |
Keywords | Field | DocType |
features induction,product named entity recognition,product named entity,random processes,information retrieval,performance improvement,domain ontology,text reviews,compound features,feature extraction,domain ontology features,chinese pner,conditional random fields,ontologies (artificial intelligence),f-measure,crf,natural language processing,external features,text analysis,internal features,semantic information,viterbi algorithm,conditional random field,f measure | Entity linking,Conditional random field,Ontology,Computer science,Named entity,Feature extraction,Natural language processing,Artificial intelligence,Named-entity recognition,CRFS,Viterbi algorithm | Conference |
ISBN | Citations | PageRank |
978-1-4673-1211-0 | 0 | 0.34 |
References | Authors | |
5 | 4 |