Beruflich Dokumente
Kultur Dokumente
Available at http://www.ijcsonline.com/
Abstract
Query processing is very important task for information retrieval. Semantic techniques are very important for efficient
information retrieval engine. Most of the existing query processing techniques mainly focused on execution time rather
than relevancy of the results. In addition, incorporation of the semantic knowledge has been limited in the query
processing techniques. In this paper, semantic knowledge base is constructed in the form of the ontology for a particular
domain. Further, the developed ontology is used to extend the query being processed. Expanded query had more
semantic knowledge as compare to initial query. Experimental results show the effectiveness of the proposed system for
query optimisation. In addition, a mapping between underlying database and developed ontology has also been
performed by normalizing the database to reduce data redundancy. 3D model is used for storing the mapped database.
3d model here is the vital part in providing intelligence to system in keyword filtering process of query processing
mechanism
Keywords: Query Processing, Ontology, Semantics.
I.
INTRODUCTION
RELATED WORK
85 | International Journal of Computer Systems, ISSN-(2394-1065), Vol. 01, Issue.03, December, 2014
Namita Mittal et al
For example:
Sentence 1: NIT Jaipur is located at Jaipur.
Here, NIT Jaipur is the subject (resource), is located at is
the predicate (property) and Jaipur is the object (propertyvalue).
C. Ontology Description
Ontology Domain concerned to the approach is Indian
Universities. By developing domain specific ontology the
flow of information is streamlined about our domain like
various web resources providing information about Indian
colleges and universities and their characteristics are
published and shared under same ontology and one can
share and aggregate information from these sources and
hence provide a more viable solution to the user query
about Indian universities/colleges. The domains of the
ontology defined can be integrated / reused by other
ontology, thus integrating several existing ontologies under
large ontology describing a large domain as described in
figure 1.
BACKGROUND
A. Ontology
Ontologies [6,7] provide a concise and univocal
description of components/classes and their relationships
for a domain of interest and this descriptive knowledge can
be used by several user and user groups. The hierarchical
organization of classes with relationships among them
constitutes the backbone of ontology.
B. Ontology Representation
The Ontologies are represented in real world through
conceptualization shared by humans as a knowledge base
and are described through machine readable languages
such as OWL (Web Ontology Language), XML(Extensible
Markup Language). Inside Ontology everything is derived
from <owl:Thing> which further comprises classes, object
type properties, data type properties and instances. The
main component of ontology is Class which is basically
description of concepts in the domain. Apart from classes
second most important component of ontology are the
properties which are classified into object properties and
datatype properties and these describes various features and
attributes of the concepts and then individual instances of
the classes are made which in together makes a knowledgebase for the system.
In the proposed approach an ontology is represented using
RDF based models [9]. Resource Descriptive Framework
(RDF) is the backbone of semantic web, which is the
standard model for data interchange on the web. It is linked
with the other resources via URI(Uniform Resource
Allocator) which uniquely identifies the documents. It
works on the XML technology hence making data in each
database understood by each other in XML form and hence
platform independent. RDF is based on statements of a
triple-form (Subject, Predicate, Object ). The meaning of a
particular RDF model is the conjunction (i.e. logical AND)
of all the statements it contains [9]. Any element of a
statement is referred to as a resource. Predicates specify
characteristics or relationships of resources, and are
referred to as properties.
86 | International Journal of Computer Systems, ISSN-(2394-1065), Vol. 01, Issue.03, December. 2014
Namita Mittal et al
Object Properties
Domain
Range
hasCourses
Universities
Courses
hasRecognizationfr
om
Universities
Recognizing
Body
isLocatedin
Universities
Zone
Domain
hasUniversityName
Universities
hasCourseName
Courses
Figure 3 Dataset-sheet1
Figure 4- Dataset-sheet2
Figure 2 Class Hierarchy
D. Dataset Description
Dataset corresponding to domain of ontology is selfdeveloped which comprises separate MS-Excel file
corresponding to different university and that is having
information of particular university about the various
classes and properties which is mentioned in our ontology.
Information gathering about various universities is done
through web resources, brochures, etc. Datasets are kept in
well normalized form so as to reduce redundancy and
proper format has been maintained for each data file and
proper keys table have also been maintained for various
many to many relationship and properties. Special care has
been given on dynamicity through a framework in regards
with further addition of database entries.
IV.
For Example:
For data about NIT Jaipur, we keep documents named NIT
Jaipur in a separate folder of NITs , which comprises of
two files, first one comprises information like Name,
University Type Zonal Location, Courses, Recognizations
where courses and recognizations may have more than one
values, hence proper keys as a value of combinations of
different values are represented into another integrated files
as shown in figure 3 and the second internal sheet
comprises of the Streams/Specialization corresponding to
87 | International Journal of Computer Systems, ISSN-(2394-1065), Vol. 01, Issue.03, December. 2014
Namita Mittal et al
={Universities,Zone=north,...,Courses=mtech}
out:
PREFIX rdfs:<http://www.w3.org/2000/01/rdf-schema#>
88 | International Journal of Computer Systems, ISSN-(2394-1065), Vol. 01, Issue.03, December. 2014
Namita Mittal et al
V.
Sample Queries
Colleges having BTech Course
in North.
Colleges for BTech, MTech and
MBA with UGC recognizations.
UGC Universities with MBBS
courses.
Universities
having
PhD
courses with UGC or MHRD
recognizations.
Universities having courses as
BTech or MBBS with UGC
recognizations.
MSc, MBBS, BTech colleges
with UGC recognization in East
zone.
Universities having MBBS
courses recognized from MCI
and are located in North or
South.
MHRD recognized north zone
universities for BTech studies.
F-score
0.956
0.963
0.896
0.965
0.984
1.0
0.986
0.952
CONCLUSION
89 | International Journal of Computer Systems, ISSN-(2394-1065), Vol. 01, Issue.03, December. 2014
Namita Mittal et al
[4]
[5]
[6]
[7]
[8]
[9]
90 | International Journal of Computer Systems, ISSN-(2394-1065), Vol. 01, Issue.03, December. 2014