Construction of the structural definition-based terminology ontology system and semantic search evaluation
| Pages | 705-732 |
| Published date | 21 November 2016 |
| DOI | https://doi.org/10.1108/LHT-08-2016-0090 |
| Date | 21 November 2016 |
| Author | Young Man Ko,Min Sun Song,Seung Jun Lee |
| Subject Matter | Library & information science,Librarianship/library management,Library technology,Information behaviour & retrieval,Information user studies,Metadata,Information & knowledge management,Information & communications technology,Internet |
Construction of the structural
definition-based terminology
ontology system and semantic
search evaluation
Young Man Ko
Department of Library and Information Science,
Sungkyunkwan University, Seoul, Republic of Korea, and
Min Sun Song and Seung Jun Lee
Institute for Knowledge and Information Management,
Sungkyunkwan University, Seoul, Republic of Korea
Abstract
Purpose –The purpose of this paper is to construct a structural definition-based terminology ontology
system that defines the meanings of academic terms on the basis of properties and links terms with
properties that are structured by conceptual categories (classes). This study also aims to test the possibility
of semantic searches by generating inference rules and setting very complicated search scenarios.
Design/methodology/approach –For the study, 55,236 keywords from the articles of the “Korea
Citation Index”were structurally defined and relationships among terms and properties were built.
Then, the authors converted the RDB data into RDF and designed ontologies using the ontology
developing tool Protégé. The authors also tested the designed ontology with the inference engine of the
Protégé editor. The generated reference rules were tested by TBox and SPARQL queries.
Findings –The authors generated inference control rules targeting high-input-ratio data in the
properties of classes by calculating the input ratio of real input data in the system, and then the authors
executed a semantic search by SPARQL query by setting very complicated search scenarios, for which
it would be difficult to deduce results via a simple keyword search. As a result, it was confirmed that
the search results show the logical combination of semantically related term data.
Practical implications –The proposed terminology ontology system was constructed with the
author keywords from research papers, it will be useful in searching the research papers which include
the keywords as search results by the complex combination of semantic relation. And the Structural
Terminology Net database could be utilized as an index database in retrieval services and the mining
of informal big data through the application of well-defined semantic concepts to each term.
Originality/value –This paper presented a methodology for supporting IR using expanded queries
based on a novel model of structural terminology-based ontology. The user who wants to access the
specific topic can create query that brings the semantically relevant information. The search results
show the logical combination of semantically related term data, which would be difficult to deduce
results via traditional IR systems.
Keywords Author keyword, Inference rule, Knowledge organization system, Semantic relationship,
Structural terminology net, Terminology ontology
Paper type Research paper
1. Introduction
1.1 The purpose of this study
Existing knowledge organization systems, such as academic glossaries or thesauruses,
struggle to capture the variety of semantic relationships between terminologies because
they simply define the terms or define onlythe broader, narrower, and related concepts.
Library Hi Tech
Vol. 34 No. 4, 2016
pp. 705-732
©Emerald Group Publis hing Limited
0737-8831
DOI 10.1108/LHT-08-2016-0090
Received 6 June 2016
Accepted 4 September 2016
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0737-8831.htm
This paper was supported by Samsung Research Fund, Sungkyunkwan University, 2014.
705
Structural
definition-
based
terminology
To overcome these problems, much research has been conducted on new knowledge
structures, such as the various ontologies based on thesauruses or the thesauruses
containing definitions of terms.
In this study, we propose a structural academic glossary as a new form of
knowledge organization system to overcome the limitations of existing knowledge
structures. The structural academic glossary described in this study defines each
academic term depending on various conceptual categories (hereafter classes) with
many properties. In the structural academic glossary, each term belonging to the same
class is defined based on the properties of that class. This study starts with the
assumption that it is possible to search semantically relevant terms efficiently if we
generate inference rules based on setting up properties, classes, and relationships about
terms through constructing a structural academic glossary database.
For the experiment, we constructed a structural academic glossary based on a
relational database system targeting author keywords of journal articles in the fields of
the humanities, social sciences, arts, and sports in the Korea Citation Index (hereafter KCI).
Theofficialnameofthissystemis“Structural Terminology Net (hearafter STNet),”and
the web address is http://stnet.re.kr. Then, we evaluated semantic search resul tsappl ying
inference rules generated by converting the RDB data of STNet into RDF ontology.
1.2 Related works
In philosophy, ontology is the study of describing the kinds of things that exist in the
world and how they are related. In information science, ontology is used to refer to a
body of knowledge describing the sorts of objects, properties of objects, and relations
between objects that are possible in a specified domain. Ontology can be applied in
many domains and a survey of Meenachi and Baba (2012) presented on the usage of
ontology in various domains like Medical, Agriculture, Geosciences, Education, Marine,
Communication, Computer, Chemical, Defence, Linguistic, etc.
Currently there are a significant number of researches to deal the issue of ontology
building methodology. The research can be divided essentially in two approaches. The
first collects terminology and builds the ontology by analyzing concepts, forming a
taxonomy for the concepts, and defining the relationships between the concepts and the
rules for acquiring domain knowledge. This work takes four directions: the bottom-up
method; the top-down method; the middle-out method; and the hybrid method. The
bottom-up method starts with specific concepts and then groups them into general
concepts (Grüninger and Fox, 1995; Van Der Vet and Mars, 1998). The top-down
method starts with the general classes and then divides these into sub-classes
(Schreiber et al., 1995). The middle-out method starts with certain mid-level concepts
and then applies the bottom-up method or the top-down method (Corcho et al., 2005;
Yoo et al., 2014). The hybrid method merges ontologies developed from the bottom-up
method and top-down method into one ontology (López-Pellicer et al., 2008).
The second approach to ontology building involves developing an ontology from
database schemas. Many methods have been reported for connecting with transferring
relational database to ontology structure (Michel et al., 2013). One of the aspects that
existing methods can be classified based on it is the type of the source of transmission.
They are roughly classified into one of the five categories: approaches based on an
analysis of relational schema (Stojanovic et al., 2002; Li et al., 2005; Sane and Shirke,
2009; Dong et al., 2013; Thuy et al., 2014), approaches based on an analysis of
tuples (Astrova, 2004; Sonia and Khan, 2008), approaches based on HTML pages
(Astrova and Stantic, 2005; Benslimane et al., 2006), approaches based on entity
706
LHT
34,4
To continue reading
Request your trial