Sei sulla pagina 1di 4

International Conference on Computing and Intelligence Systems

Pages: 1387 1390

Volume: 04, Special Issue: March 2015


ISSN: 2278-2397

Semantic Information Retrieval Based on


Domain Ontology
A. Sudha Ramkumar1, B. Poorna2
1

Research Scholar, Bharathiar University , Coimbatore, India,


2
Principal, SSS Jain College,Chennai, India.
Email: sudharam99@gmail.compoornasundar@yahoo.com

Abstract - Most of the current search engines


are based on the keyword search and it returns a
large amount of irrelevant information in response
to the user query. This is because of the keyword
matching does not consider the semantics of the
keyword term. Ontology has been increasingly
used to efficiently capture the semantics of the
information which overcomes the problems
associated with traditional keyword search. The
importance of ontology is identified in the area of
building semantic information retrieval systems
for the specific domain and gains greater
significance
over the tradition keyword
search.Knowledge representation formalism such
as description logic is used to obtain the knowledge
of the domain of interest and is used for the
development of intelligent information retrieval
system.Protg is the widely used tool for
developing and editing ontology with GUI which
enables the ontology developers to describe the
conceptual terms and relationship between them
by creating taxonomies. This paper describes the
steps for creating operating system ontology and
the various aspects like subclassof, superclassof
relationships and an overview of search algorithm
have been described.
Keywords - Semantic Information retrieval;
Search algorithm; Ontology.
I. INTRODUCTION
Traditional keyword search relies on the
occurrence of the keywords in the document and this
search returns a large amount of irrelevant
information.Knowledge management is the process of
capturing, structuring and reusing the information to
develop an understanding of how a particular system
works, and subsequently to share this information
meaningfully to other information system. The
adequate management of knowledge is becoming
more important for academicians in educational
institutions. Ontology is used to represent the
knowledge of the domain of interest which
significantly improves the search results. According
to Gruber ontology is an explicit specification of
conceptualizations. Ontologies are developed with the
help of ontology languages such as web ontology
language OWL, Resource Description Framework

International Journal of Computing Algorithm (IJCOA)

RDF, and RDF-schema. Ontologies serve to provide


conceptualization in the domain of interest and thus
guarantee to provide semantics of the information.
Thus ontologies are an integral part in every
knowledge management initiative. The protg tool
developed by the Stanford university is widely used
tool for constructing domain ontology and to develop
semantic information retrieval system.
This paper provides an overview of development
of ontology and the use of DL query. Finally we
discuss how a DL queries are used for the semantic
search. This paper is organized as follows. Section 2
provides the overview of semantic information
retrieval systems, ontology languages and DL queries.
Section 3 describes our framework of semantic
information retrieval system based on ontology.
Sections 4 provides an overview of search algorithm
and section 5 describes the future work and
conclusion.
II. BACKGROUND
The semantic web aims to extend the current web
standards and technology so that the semantics of the
web is machine processible[12] . The use of domain
ontology in the semantic information retrieval system
enablesusers torepresent a meaningful knowledge of a
particular domain of interest which makes the
machine more intelligent to understand the semantics
of the information.Domain Ontology describes a list
of terms to represent important concepts, such as
classes of objects and the relationship between them
in order to represent the knowledge of domain[13].
Gruber defines ontology as an explicit specification
of conceptualization. Ontologies play an important
role in representing the terms unambiguously which
serve as a background for machine processible.
Ontology languages are formal languages and are
used to construct ontologies. They allow the encoding
of knowledge about the specific domain and often
include reasoning rules that support the processing of
the available knowledge. Ontology languages are
commonly based on either first order logic or on
description logic. Among the several ontology
languages
such
as
Resource
Description
Framework(RDF),
RDF-Schema,Web
ontology
language OWL, OWL is the most acceptable and
recommended ontology language by W3C to

1387

International Conference on Computing and Intelligence Systems


Pages: 1387 1390
construct a domain ontology[5].Sir Jorge Cardoso[2]
surveyed on most widely used ontology editors and
found that protg tool had a market share of 68.2%
followed by swoop, ontoedit, texteditor, altova
semantic works and hence ontologies were mostly
developed in the educational field 31%.
Semantic information retrieval using ontology in
university domain[10] developed and implemented a
search engine confined to the university domain where
the query is analysed semantically to obtain the
accurate result. Semantic search using ontology and
RDBMS for cricket[9] uses reference to ontology data
directly from SQL using semantic match operators and
focuses on semantic search based on ontology.
Developing an university ontology in education
domain using protg for Semantic web[6]by Sanjay
Kumar Maliket al., demonstrated the indraprastha
university ontology with aspects like super class,
subclass, query retrieval process and TGVIZ graph
visualization.
Developing university ontology using protg owl
tool: Process and Reasoning[7] by Naveen Malviya et
al., is similar to Sanjay Kumar malik et al., .The Jena
based ontology model inference and retrieval
application[14] by Luo Zhong et al., proposed
computer graphics ontology model and stored the owl
documents in MySQL, sets the inference rule and
achieve search queries on the ontology database.
Semantic Search : Document ranking and clustering
using computer science ontology and N-grams[1] by
Thanyaporn Boonyoung presented a new document
ranking process that proposes a new weight of query
term in the document based on the computer science
ontology.
III.

FRAMEWORK OF PROPOSED SYSTEM

The operating system ontology has been


developed using protg 4.2 with the related classes
as conc epts. All the concepts related to the operating
system ontology has been defined in the hierarchy
with data properties, object properties and annotation
properties. The object property defines the
relationship between the concepts of operating system
ontology and the data properties used to define the
relationship between the concept individual and the
data literal. The annotation properties such as
comment, seealso, isdefinedby etc are used to
annotate the concepts of the domain ontology of
interest. Definition of each class, subclass concept
and list of references to each class and subclass are
attached to the node of class hierarchy. The ontoGraf
tab is used to visualize the ontology concepts like
graph. The DL query tab of protg is a powerful and
easy to use feature for searching a classified ontology.
This query language is based on the OWL syntax that
is fundamentally based on collecting all the
information about a class, property or instances into a
single frame.

International Journal of Computing Algorithm (IJCOA)

Volume: 04, Special Issue: March 2015


ISSN: 2278-2397

Fig 1: Framework of Proposed system

Fig 2: Class Hierarchy of OS ontology

Our proposed system first identifies the concepts


and their relationship for the operating system
ontology. The identified concepts and their
relationship are built into domain ontology and stored
in the knowledge repository with sufficient annotation
properties and instances. Finally, this ontology is
called through the user interface by matching the
keyword term with the concepts present in the class
hierarchy. The search algorithm is to be used to search
the concepts of this domain ontology and to obtain tha
knowledge of that concept.

Fig 3: Visualization of ontology

1388

International Conference on Computing and Intelligence Systems


Pages: 1387 1390

Volume: 04, Special Issue: March 2015


ISSN: 2278-2397

spreading activation[8] is used with the query


expansion technique will yield very good search
results.
V. CONCLUSION AND FUTURE WORK

Fig 4: DL query

IV. OVERVIEW OF SEARCH ALGORITHM


Hildebrand[4] examined 35 search systems and
according to him, the search system consists of 3 main
steps. They areQuery construction, Core search
process and presentation and organization of
results.The core search process contains the search
algorithm. All keyword based search involves some
form of syntactic matching of the query against textual
content or metadata. The semantic matching extends
the syntactic matching by exploiting the linked data in
the semantic graph. To index the textual content of the
ontology, a search engine such as Lucene can be used
along with a triple store.
When writing search algorithm, the easiest way to
optimize code is to create inverted index which is a
structure which maps each searchable keyword to a
list of document that occurs in. Along with each
document, store a corresponding score for that
keyword in that document. To calculate score, there
are many search algorithms[4] available like weighted
graph search algorithm may constrain the possible
path structures and path length, e-culture algorithm
assigns weights manually to RDF relations, SemRank
automatically computes weights based on statistics
derived from graph structures, and spreading
activation algorithm[11] incorporates weights as well
as the number of incoming links. The vector space
model uses the spreading activation algorithm by
exploiting the relationship between the concepts will
give the required appropriate information to the user in
response to their query.
The spreading activation model[3] is widely used
in the information retrieval which consists of a
network structure of concepts of a domain and
processing techniques applied on this structure to find
the concepts related to the query term. The relations
between concepts of a domain are usually labelled and
weighted. The spreading activation algorithm will
create a set of concepts after receiving the contents of
the query. With this set of concept, it will find the
nearest concepts in the network structure. This process
is an iterative one. This spreading activationalgorithm
is like waves activates from one concept to all other
conceptsconncected to the network. When this

International Journal of Computing Algorithm (IJCOA)

The ontology becomes an integral part of


every semantic applications. The use of ontology in
the search process will make the machine
understandable about the information and helps to
produce intelligent search results. The use of ontology
in the spreading activation algorithm increases the
coverage of relationship between the concepts. This
paper developed a operating system ontology in
protg along with relationships, check for
consistency of classes, visualizates the ontology and
finally describes the usage of DL query on ontology
concepts. The scope of this developed operating
system is very small. Our future work is to call the
ontology concepts as network structure consists of the
nodes and arcs, here the concepts are represented as
nodes and properties are represented as arcs. This
network structure is to be applied to the spreading
activation algortithm to find relevant information
based on the result of traditional keyword search based
on the document index.
REFERENCES
[1]

Boonyoung, Thanyaporn, and Anirach Mingkhwan. "Semantic


Search Using Computer Science Ontology Based on Edge
Counting and N-Grams." Recent Advances in Information and
Communication
Technology.
Springer
International
Publishing, 2014. 283-291.
[2] Cardoso, Jorge. "The semantic web vision: Where are
we?." Intelligent Systems, IEEE 22.5 (2007): 84-88.
[3] Crestani, Fabio. "Application of spreading activation
techniques in information retrieval." Artificial Intelligence
Review 11.6 (1997): 453-482.
[4] Hildebrand, Michiel, Jacco Van Ossenbruggen, and Lynda
Hardman. "An analysis of search-based user interaction on the
semantic web." (2007).
[5] Horridge, Matthew, et al. "A Practical Guide To Building
OWL Ontologies Using The Protg-OWL Plugin and COODE Tools Edition 1.0." University of Manchester (2004).
[6] Malik, Sanjay Kumar, Nupur Prakash, and S. A. M. Rizvi.
"Developing an university ontology in education domain using
protg for semantic web."International Journal of Science and
Technology 2.9 (2010): 4673-4681.
[7] Malviya, Naveen, Nishchol Mishra, and Santosh Sahu.
"Developing university ontology using protg owl tool:
Process and reasoning." International Journal of Scientific &
Engineering Research 2.9 (2011): 1-8.
[8] Ngo, Vuong M. "Discovering Latent Informaion By Spreading
Activation Algorithm For Document Retrieval." International
Journal of Artificial Intelligence & Applications 5.1 (2014).
[9] Patil, S. M., and D. M. Jadhav. "Semantic Search using
Ontology and RDBMS for Cricket." International Journal of
Computer Applications 46 (2012).
[10] Rajasurya, Swathi, et al. "Semantic information retrieval using
ontology
in
university
domain." arXiv
preprint
arXiv:1207.5745 (2012).
[11] Schumacher,
Kinga,
Michael
Sintek,
and
Leo
Sauermann. Combining fact and document retrieval with
spreading activation for semantic desktop search. Springer
Berlin Heidelberg, 2008.

1389

International Conference on Computing and Intelligence Systems


Pages: 1387 1390

Volume: 04, Special Issue: March 2015


ISSN: 2278-2397

[12] Shadbolt, Nigel, Wendy Hall, and Tim Berners-Lee. "The


semantic web revisited." Intelligent Systems, IEEE 21.3
(2006): 96-101.
[13] Swartout, Bill, et al. "Toward distributed use of large-scale
ontologies." Proc. of the Tenth Workshop on Knowledge
Acquisition for Knowledge-Based Systems. 1996.
[14] Zhong, Luo, et al. "The Jena-Based Ontology Model Inference
andRetrieval
Application." Intelligent
Information
Management 4.04 (2012): 157.

International Journal of Computing Algorithm (IJCOA)

1390

Potrebbero piacerti anche