Research on domain ontology in different granulations based on concept lattice
Xiangping Kang
a,b
, Deyu Li
a,b,
⇑
, Suge Wang
a,b,c
a
School of Computer and Information Technology, Shanxi University, Taiyuan, 030006 Shanxi, China
b
Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education, Taiyuan, 030006 Shanxi, China
c
School of Mathematics Science, Shanxi University, Taiyuan, 030006 Shanxi, China
article info
Article history:
Received 18 November 2010
Received in revised form 20 September 2011
Accepted 21 September 2011
Available online 29 September 2011
Keywords:
Concept lattice
Granular computing
Ontology building
Ontology merging
Ontology connection
abstract
This paper introduces concept lattice and granular computing into ontology learning, and presents a uni-
fied research model for ontology building, ontology merging and ontology connection based on the
domain ontology base in different granulations. In this model, as the knowledge in the lowest and most
basic level, the domain ontology base is presented firstly, which provides a uniform technology for ontol-
ogy learning on the whole; secondly, in order to better understand problems rather than be overwhelmed
unnecessary details, granular computing is introduced to abstract and simplify domain ontology bases in
complex domains. Moreover, the similarly of concepts in different granulations is introduced to help
domain experts judging relations except for inheritance relation, and the similarity of ontologies in
multi-granulations is introduced to measure the degree of connection of ontologies; finally, based on
similarity models mentioned above, the ontology building, ontology merging and ontology connection
can be obtained in different granulations with the help of domain experts. It is shown by instances that
the application of the model presented in this paper is valid and practicable. Although there are still some
problems in applications of this model (for example, ontology learning cannot dispense with the inter-
vention of domain experts yet), this paper offers a new way for combining ontology learning and concept
lattice.
Crown Copyright Ó 2011 Published by Elsevier B.V. All rights reserved.
1. Introduction
The term ontology is borrowed from philosophy. In 1993
Gruber proposed a most popular definition of ontology, that is,
ontology is an explicit specification of a conceptual model [1]. Later
Borst modified it slightly, and he presented that ontology is a for-
mal and explicit specification of a shared conceptual model [2].
Although ontology is defined in different ways, researchers’ recog-
nition of it is unified from the perspective of essence. Since 1990s,
research, exploitation and applications of ontology in computer
area has become a hot issue, and ontology has gradually attract
much attention of researchers in many areas, such as knowledge
acquisition and representation, planning, process management,
database framework integrated, natural language processing, busi-
ness simulation etc. Along with going deeper into applications of
ontology, some new mature and effective ontology learning meth-
ods are needed to support new demands, so it is necessary to make
an active exploration on new methods. In recent years, many tools
and techniques have been used to ontology research, and then a
mass of theories and methods have been achieved [3–11], such
as the use of formal concept analysis cannot only weaken subjec-
tive effects in the process of ontology learning from developers,
but also can automatically acquire implied concepts and relation-
ships among concepts.
Concept consists of extent and intent, and based on this philos-
ophy Professor Wille [12] proposed formal concept analysis (FCA),
which is a method for finding, ordering and displaying concepts in
early 1980s. Concept lattice is an effective tool in FCA, and it is very
suitable for mining potential concepts of dates. It has been widely
studied [14–20] and applied to machine learning [21], software
engineering [22] and information retrieval [23].
Ontology aims to build a shared model for the objective world
perceived by human, but concept lattice builds the model for arti-
ficial world rather than real world. We can build ontology for a cer-
tain domain when there is no data, but concept lattice must be
built on a given data set. From the view of ISO704 standard shown
in Fig. 1, it is not hard to see that the focus concerned by concept
lattice and ontology is different. Concept lattice pays its concerns
on the concept level, while ontology more concerns the presenta-
tion level. These two formal methods of ontology and concept lat-
tice have little difference, but concept lattice viewed as a useful
tool can be introduced into ontology. There are some common
application fields of concept lattice and ontology in the philosophy,
0950-7051/$ - see front matter Crown Copyright Ó 2011 Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.knosys.2011.09.016
⇑
Corresponding author at: School of Computer and Information Technology,
Shanxi University, Taiyuan, 030006 Shanxi, China.
E-mail addresses: kangxiangping@yahoo.cn (X. Kang), lidysxu@yahoo.cn (D. Li),
wsg@sxu.edu.cn (S. Wang).
Knowledge-Based Systems 27 (2012) 152–161
Contents lists available at SciVerse ScienceDirect
Knowledge-Based Systems
journal homepage: www.elsevier.com/locate/knosys