TY - GEN
T1 - A Wikipedia based hybrid ranking method for taxonomic relation extraction
AU - Zhong, Xiaoshi
PY - 2013
Y1 - 2013
N2 - 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.
AB - 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.
KW - Wikipedia
KW - hybrid ranking method
KW - select best position
KW - taxonomic relation extraction
UR - http://www.scopus.com/inward/record.url?scp=84893286464&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-45068-6_29
DO - 10.1007/978-3-642-45068-6_29
M3 - Conference contribution
AN - SCOPUS:84893286464
SN - 9783642450679
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 332
EP - 343
BT - Information Retrieval Technology - 9th Asia Information Retrieval Societies Conference, AIRS 2013, Proceedings
T2 - 9th Asia Information Retrieval Societies Conference on Information Retrieval Technology, AIRS 2013
Y2 - 9 December 2013 through 11 December 2013
ER -