A Wikipedia based hybrid ranking method for taxonomic relation extraction
|A Wikipedia based hybrid ranking method for taxonomic relation extraction|
|Published in||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Keyword(s)||hybrid ranking method, select best position, taxonomic relation extraction, Wikipedia (Extra: Best position, Inference methods, Ranking methods, Relation extraction, Wikipedia, Extraction, Information retrieval, Infrared devices, Taxonomies)|
|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|
A Wikipedia based hybrid ranking method for taxonomic relation extraction is a 2013 conference paper written in English by Zhong X. and published in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics).
This paper proposes a hybrid ranking method for taxonomic relation extraction (or select best position) in an existing taxonomy. This method is capable of effectively combining two resources, an existing taxonomy and Wikipedia, in order to select a most appropriate position for a term candidate in the existing taxonomy. Previous methods mainly focus on complex inference methods to select the best position among all the possible position in the taxonomy. In contrast, our algorithm, a simple but effective one, leverage two kinds of information, the expression of and the ranking information of a term candidate, to select the best position for the term candidate (the hypernym of the term candidate in the existing taxonomy). We conduct our approach on the agricultural domain and the experimental result indicates that the performances are significantly improved.
- This section requires expansion. Please, help!
Probably, this publication is cited by others, but there are no articles available for them in WikiPapers.