BioKeySpotter: An Unsupervised Keyphrase Extraction Technique in the Biomedical Full-Text Collection
Document Type
Article
Publication Date
12-1-2012
Abstract
Extracting keyphrases from full-text is a daunting task in that many different concepts and themes are intertwined and extensive term variations exist in full-text. In this chapter, we proposes a novel unsupervised keyphrase extraction system, BioKeySpotter, which incorporates lexical syntactic features to weigh candidate keyphrases. The main contribution of our study is that BioKeySpotter is an innovative approach for combining Natural Language Processing (NLP), information extraction, and integration techniques into extracting keyphrases from full-text. The results of the experiment demonstrate that BioKeySpotter generates a higher performance, in terms of accuracy, compared to other supervised learning algorithms. © Springer-Verlag Berlin Heidelberg 2011.
Identifier
84885630694 (Scopus)
ISBN
[9783642231506]
Publication Title
Intelligent Systems Reference Library
External Full Text Location
https://doi.org/10.1007/978-3-642-23151-3_3
e-ISSN
18684408
ISSN
18684394
First Page
19
Last Page
27
Volume
25
Recommended Citation
Song, Min and Tanapaisankit, Prat, "BioKeySpotter: An Unsupervised Keyphrase Extraction Technique in the Biomedical Full-Text Collection" (2012). Faculty Publications. 17893.
https://digitalcommons.njit.edu/fac_pubs/17893
