A Survey on Ontology Matching Techniques
Satish Kumar
National Institute of Technology,
Kurukshetra, India.
er.satishkumar149@gmail.com
Vikram Singh
National Institute of Technology,
Kurukshetra, India.
viksnitkkr@outlook.com
Balwinder Saini
National Institute of Technology,
Kurukshetra, India.
me7saini@gmail.com
Abstract— Words, sentences, paragraphs and documents is an
important component in the data mining application area such
as information retrieval, text summarization etc. And finding
the similarity value between these components is an important
action. Now a day, there are various matching techniques are
present which are very helpful for finding the similarity
between the ontology. In this survey paper discusses the
existing works on similarity (On text, words etc.) using
element-level techniques and structure level techniques.
Index Terms— Similarity, String-Based similarity, Element-
Level techniques, Structure-Level techniques
I. INTRODUCTION
Homogeneous attribute measures play a consequential role in
text cognate area and application in tasks such as ontology
matching, information retrieval, query answering etc [1, 2].
Measuring the kindred attribute between the words, text,
documents is the primary stage in ontology matching.
Ontology matching is a paramount operation traditional
applications e.g., Ontology evolution, Ontology integration,
data integration and data warehouses [2]. In text similarity
words can be similar in two ways lexically and semantically.
If the sequence of character is similar then they are lexically
similar and if the two words are opposite of each other having
similar meaning then they are semantically similar [3]. These
types of similarities measures through the element-level
techniques between two ontologies. Element –Level
techniques discuses the String-predicated techniques,
Language-predicated techniques and Constraint-Based
techniques etc.
On the other hand Structure-Level techniques are used to
measure the similarity based on the structure. Structure-Level
techniques discuss the Graph-predicated technique, Taxonomy-
predicated technique and Constraint-predicated technique.
II. MATCHING PROCESS
The ontology matching process is utilized to measures the
homogeneous attribute between the two ontologies [4]. The
matching process determines an alignment A', for two
ontologies O
1
and O
2
. There are some other parameters of the
matching process.
(i) The utilization of an input alignment A
(ii) Matching parameter like weights, threshold
(iii) External resources utilized by the matching process like
erudition-base and domain-categorical area.
Definition 1 (Matching process) The matching process
considered as function f which emanates from the ontologies
to match O
1
and O
2
, an input alignment A, a set of parameters
p and a set of oracles and resources r, back to an alignment A'
between these ontologies[1, 4].
A' = f (O
1
, O
2
, A, p, r)
Fig. 1 The Matching Process
III. TYPES OF ONTOLOGY MATCHING TECHNIQUES
There are two types of otology matching techniques discusses.
First one is Element-level technique and another one is
Structure-level techniques [4].
Fig. 2 Types of Ontology Matching Techniques
13
2014 5th International Conference on Computer and Communication Technology (ICCCT)
987-1-4799-6756-8/14/$31.00©2014 IEEE