In this paper, a method is proposed to discriminate acronyms and their full names or expansions in scientific and technical literature abstracts by learning Wikipedia definition statements. Through this study, we aim to verify the effective utilization of an open knowledge base in the knowledge processing of domain-specific fields. Experimental results confirm that a noun phrase (NP)-type feature has better performance than a noun (NN) type feature in terms of precision rate. On the contrary, the results of measuring query response rate indicate that a single NN-type feature has better performance than an NP-type feature. We also verify that additional collocation information can contribute to improve the response rate. This study is mainly divided into three parts: 1) a process of sense discrimination is classified into many steps according to feature types; 2) the measured results are combined and processed; and 3) a data fusion-based incremental approach is proposed for sense discrimination. Through the method, we can adjust a precision rate to a certain level while considering classifier response rate.
[1]
Jens Lehmann,et al.
DBpedia: A Nucleus for a Web of Open Data
,
2007,
ISWC/ASWC.
[2]
Won-Kyung Sung,et al.
Generating Knowledge Map for Acronym-Expansion Recognition
,
2011,
FGIT-UNESST.
[3]
Angela Fogarolli,et al.
Word Sense Disambiguation Based on Wikipedia Link Structure
,
2009,
2009 IEEE International Conference on Semantic Computing.
[4]
T. Landauer,et al.
Indexing by Latent Semantic Analysis
,
1990
.
[5]
Robert P. Cook,et al.
Freebase: A Shared Database of Structured General Human Knowledge
,
2007,
AAAI.
[6]
Hanmin Jung,et al.
Comparative study on disambiguating acronyms in the scientific papers using the open knowledge base
,
2014,
ICIST 2014.