Title named entity recognition using wikipedia and abbreviation generation
|Title named entity recognition using wikipedia and abbreviation generation|
|Author(s)||Park Y., Kang S., Seo J.|
|Published in||2014 International Conference on Big Data and Smart Computing, BIGCOMP 2014|
|Keyword(s)||Abbreviation generation, Conditional random field, Title named entity, Wikipedia (Extra: Abbreviation generation, Additional costs, Conditional random field, Lexical information, Named entities, Named entity recognition, Statistical modeling, Wikipedia, Image segmentation, Random processes, Natural language processing systems)|
|Article||BASE, CiteSeerX, Google Scholar|
|Web||Ask, Bing, Google (PDF), Yahoo!|
|Download and mirrors|
|Local copy||Not available|
|Remote mirror(s)||Not available|
|Export and share|
|BibTeX, CSV, RDF, JSON|
|Browse properties · List of conference papers|
Title named entity recognition using wikipedia and abbreviation generation is a 2014 conference paper written in English by Park Y., Kang S., Seo J. and published in 2014 International Conference on Big Data and Smart Computing, BIGCOMP 2014.
In this paper, we propose a title named entity recognition model using Wikipedia and abbreviation generation. The proposed title named entity recognition model automatically extracts title named entities from Wikipedia so constant renewal is possible without additional costs. Also, in order to establish a dictionary of title named entity abbreviations, generation rules are used to generate abbreviation candidates and abbreviations are selected through web search methods. In this paper, we propose a statistical model that recognizes title named entities using CRFs (Conditional Random Fields). The proposed model uses lexical information, a named entity dictionary, and an abbreviation dictionary, and provides title named entity recognition performance of 82.1% according to experimental results.
- This section requires expansion. Please, help!
Probably, this publication is cited by others, but there are no articles available for them in WikiPapers.