liesp.insa‐lyon.fr
1
1
A Semantic Approach
for Classification of Web
Ontologies
M. Fahad, N. Moalla, A. Bouras
liesp.insa‐lyon.fr
2
Outlines
Introduction
Related Work
Ontology Classification Problem
Semantic based Ontology Classifire
Preliminary Results/Experiment
Lessons to Learn and Conclusion
liesp.insa‐lyon.fr
3
Outlines
Introduction
Related Work
Ontology Classification Problem
Semantic based Ontology Classifire
Preliminary Results/Experiment
Lessons to Learn and Conclusion
3
liesp.insa‐lyon.fr
•Virtual communities on Semantic Web
•Notion of ontologies
–conceptualization and elicitation of the domain
knowledge
–in a machine understandable and processable
manner
•Due to their capacities of decidability and
expressiveness, ontologies have played a fundamental
role for describing semantics of data every where, for
Information storage, processing, retrieval, decision
making
•Significant no of online ontologies raise new challenges
4
Introduction
liesp.insa‐lyon.fr
•Searching the relevant knowledge is one of the main
problems for the current and the emerging semantic web and
ontology based knowledge management business
applications, too.
•This requires proper classification of the web ontologies that
is essential to many tasks, such as:
–development of ontologies directories on web [Dmoz, 07],
–focused crawling for ontology retrieval [Ehrig, 05],
–improving quality of search [Pan, 06],
–concept specific modular ontology analysis [Seidenberg,
06]
5
Introduction
liesp.insa‐lyon.fr
•Classificationis traditionally defined as a
supervised learning problem in which a set of
labelled data is used to train a classifier that
can be used to label future examples
[Mitchell, 97].
•Ontology Classificationis a challenging
problem for efficient and effective ontology
management and retrieval for the semantic
web and enterprise ontology based business
applications.
6
Introduction
liesp.insa‐lyon.fr
•Prior to Ontology classification, much work has been done
for web page classification that aims at assigning a web
page to one or more predefined category labels
[Chakrabarti, 02].
•The current web is a heterogeneous infrastructure
containing unstructured or semi‐structured data of various
types. This opens up other number of classification
research problems:
–web site classification [Peng, 02; Glover, 02],
–blog classification [Qu, 06],
–image classification [Bosch, 07],
–semantic web page classification, etc.
7
Introduction
liesp.insa‐lyon.fr
•In recent years, many semantic web portals are
developed to facilitate ontology searching, ranking
and classification.
•But, these existing approaches exploit keywords,
phrases and termsabout ontologies rather than the
semantic knowledge hidden within the structureof
ontologies for their classification.
•The consequence is that the semantic of
information knowledge is not understandable by
machine and become a bottleneckin the process of
ontology searching and retrieval on the Web.
8
Introduction
liesp.insa‐lyon.fr
•This requires new approaches for ontology classification
based on structural knowledge and semantics to meet the
requirements and core challenges for current landscape of
ontology based research.
•Thus, our main idea behind this work is to replace the plain
text classificationalgorithm in the process of ontology
classification with ontology specificclassification algorithm.
•The proposed approach uses category ontologyrather than
bag‐of‐words for classification of arbitrary ontology by
analysing structure of knowledge hidden in ontologies.
9
Introduction
liesp.insa‐lyon.fr
10
Outlines
Introduction
Related Work
Ontology Classification Problem
Semantic based Ontology Classifire
Preliminary Results/Experiment
Lessons to Learn and Conclusion
10
liesp.insa‐lyon.fr
•There are many applications that make use of
ontologies for the classification of web documents,
emails, text categorization and many other tasks for
knowledge management and retrieval.
•Grobelnik and Mladenic (2005) report
–classification of Web documents into large topic ontology
–exploiting content of the document to be classified
–information on the web page context which is obtained
from the link structure of the Web
11
Related Work
liesp.insa‐lyon.fr
•Taghva et al. (2003) propose
–ontology‐based system for classification of emails,
–employing ontology that is later on applies rules for
identification of features for classification of emails.
–From the training set of emails, associated probabilities for
features are calculated and used as a part of the feature
vectors for an underlying Bayesian classifier.
•Wu et al. (2003) describe
–A methodology for ontology‐based text categorization,
–In which the domain ontologies are automatically acquired
through morphological rules and statistical methods.
12
Related Work
liesp.insa‐lyon.fr
•The role of ontologies is magical in classifying objects in various
enterprise applications and improving system‘s overall
performance
•But, ontology classification itself is a problemwhich should be
addressed in a semantic way for its own efficient management
and retrieval for emerging semantic web.
•Learn from the experirences of current web, if we dont focus on
key issue than emerging web would also suffer
•With the passage of time semantic web is gaining much
popularity and hence there is a significant growth seen in the
ontology development and reuse.
•This increases the demands for searching of the relevant domain
ontologies over the web.
13
Related Work
liesp.insa‐lyon.fr
•Ontologies, especially those developed in OWL
–Significantly complex data structures than mere web pages,
–OWL builds up several levels of complexity on top of the XML of
conventional web data.
–Constructs taxonomy, properties, relations, axiomatic
definitions, etc..
•Moreover by defining terms on similar or the same concepts, often
these ontologies overlap with each other.
•For example, as mentioned in one of the research studies for the
development of web portal, Swoogle, searches over 300distinct
terms that appear to stand for the only ‘Person’concept [Ding, 05].
•It is likely that large and complex ontologies will require a novel
solution and a central index of ontologies for fulfillment of sound
semantic web vision.
14
Related Work
liesp.insa‐lyon.fr
•Ontokhoj, a semantic web portal. [Patel, 03; Supekar, 03]
–allows engineers and software agents to retrieve trustworthy
ontologies,
–expedite the process of ontology engineering through extensive reuse
of ontologies
–exploits and extends the strategy of ranking based on citationsas
used by Google PageRank,
–uses semantic crawling technique to search and retrieve ontologies.
–treats ontology as plain text and uses text classification algorithms for
ontology classification
–which is the biggest drawback for ontology classification especially for
overlapping ontologies because plain text classification algorithms
only use keywords that results in poor performance and hence
classifier’s accuracy compromises.
15
Related Work
liesp.insa‐lyon.fr
•Swoogle, a semantic web search engine. [Ding, 05]
–based on metadata engine and retrieval system for the
semantic web
–makes use of multiple crawlers to find semantic web
documents and ontologies by meta‐search.
–does not provide classification mechanism and does not
use ontology context specific search for finding ontologies.
16
Related Work
liesp.insa‐lyon.fr
•Ontosearch2 , a ontology search engine. [Pan, 06]
–developed to address the problem of finding ontologies
appropriate for desired domains.
–makes use of the semantic entailments for searching
rather
than only using keywords or metadata like
SWOOGLE and Ontokhoj.
–provides restricted query interface by keyword search
only, and for ranking they use the citations to an ontology
or links to an object
within
the
Abox (assertional box) of
ontology.
–provides Tbox (terminological box) searching facility, Abox
searching mechanism and other
search directives by
allowing these restrictions
on the
search
query, and
performs the search on the desired portion.
17
Related Work
liesp.insa‐lyon.fr
•Ontolingua, ontology server. [Ontolingua, 10]
–facilitates us a distributed collaborative
environment to browse, edit, create, modify and
use ontologies
–requires user to first gets registered and then
perform the desired piece of work
18
Related Work
liesp.insa‐lyon.fr
19
Outlines
Introduction
Related Work
Ontology Classification Problem
Semantic based Ontology Classifire
Preliminary Results/Experiment
Lessons to Learn and Conclusion
19
liesp.insa‐lyon.fr
•Ontologies, especially that are developed in OWL,
–are more than texts
–contain a lot of structural and context information
–in terms of classes, datatype properties, object
properties, parent‐child relationships, description
logic (DL) axioms, individuals, etc.
•Therefore, plain text classification algorithms that
benefit the web page classification are not much
useful for ontology classification and searching
on the semantic web.
20
Ontology Classification Problem
liesp.insa‐lyon.fr
•Hence, few recent developments are seen in literature for
meeting the current challenges of semantic web.
•However, very little work has been done specifically for
ontology classification, so first we define specific terms of
ontology classification for promoting understandability based
on terminology used in web page classification.
•The general problem of ontology classification can be divided
into more specific problems depending upon
–the number of classes in the problem of interest,
–domain knowledge modeled in ontologies,
–and number of classes that can be assigned to ontology
instance, etc.
21
Ontology Classification Problem
liesp.insa‐lyon.fr
•Based on the number of classes in the problem,
classification can be divided into binary, ternary, or
multiclass ontology classification.
–Binary ontology classification
•Categorizes instance ontologies into exactly one of two
classes.
–Multiclass ontology classification
•Associates instance ontologies with more than two
classes or categories.
22
Types of Ontology Classification
liesp.insa‐lyon.fr
23
Ontology Repository
Binary Classification
RDF
OWL
Example of Binary Classification
liesp.insa‐lyon.fr
•Based on the number of classes that can be assigned
to instance ontology, classification can be divided
into single‐label and multi‐label ontology
classification.
–Single‐label strategy
•deals with assigning one and only one class label to
each instance ontology
–Multi‐label strategy
•deals with assigning more than one class to instance
ontology
24
Types of Ontology Classification
liesp.insa‐lyon.fr
•Based on the type of class assignment,
classification can be divided into hard or soft
ontology classification.
–Hard ontology classification
•determines whether an instance can either be or not
be in a particular class.
–Soft ontology classification
•predicts an instance to be in some class with some
likelihood and often a probability distribution across all
classes.
25
Types of Ontology Classification
liesp.insa‐lyon.fr
26
Ontology
Repository
Sports
Music
B
usiness
Arts
Multi
‐
Class
and
Single
‐
Label
Hard
Classification
Example
of
Multi
‐
Class
and
Single
‐
Label
Hard
Classification
liesp.insa‐lyon.fr
•Based on the organization of categories, ontology
classification can be taken as flat classification
scheme or hierarchical classification.
–Flat ontology classification
•deals with the categories that are considered parallel.
–Hierarchical ontology classification
•deals with the categories that are organized in a hierarchical
tree‐like structure (e.g., RDF(S) ontology), in which each
category may have a number of subcategories.
27
Types of Ontology Classification
liesp.insa‐lyon.fr
28
Example of Flat Classification
Arts
Sport
Movie
Magazine
Music
Flat Classification
28
liesp.insa‐lyon.fr
29
Example of Hierarchical Classification
29
Rock
29
Arts
Sport
Movie
Magazine
Hierarchical
Classification
Metal
Pop
Classical
Soft
Music
liesp.insa‐lyon.fr
30
Example of Hierarchical Classification
30
Rock
30
Arts
Sport
Movie
Magazine
Hierarchical
Classification
Metal
Pop
Classical
Soft
Music
Le
v
e
l
1
Le
v
e
l
1
Le
v
e
l
2
Le
v
e
l
2
liesp.insa‐lyon.fr
•Based on the domain knowledge modeled in ontologies,
classification can be divided into
subject, functional and
sentimental ontology
classification.
–Subject ontology classification
•categorizes ontologies depending on what is the domain and topicof
ontologies, e.g., art, disease, business, sports, etc.
–Functional ontology classification
•determines the role that the ontology plays, e.g., admission ontology,
personal home page ontology, patient examination ontology, etc.
–Sentimental ontology classification
•determines the messages or opinion that is presented in ontologies,
e.g., message between business processes or stock exchange
conditions, interaction between multi‐vendor semantic systems,
author’s attitude in blog ontology, etc.
31
Types of Ontology Classification
liesp.insa‐lyon.fr
32
Outlines
Introduction
Related Work
Ontology Classification Problem
Semantic based Ontology Classifire
Preliminary Results/Experiment
Lessons to Learn and Conclusion
32
liesp.insa‐lyon.fr
•A semantic based ontology classifier, ONTCLASSIFIRE, aims at
classifying ontologies in one or more predefined categoriesfor
efficient ontology management and search.
•First, there is a category ontologycontaining all the predefined
categories as concepts
•Second, for each predefined category, there is a representative
domain ontologyrather than bag of words for classification.
•ONTCLASSIFIREmatches domain ontology with the arbitrary
ontologiesand calculates the match rank.
•To meet the needs, we adopted a soft classification approach,
where instance ontology is predicted to be in some class with
some likelihood with a probability distribution across all classes.
33
A Semantic approach for Classification of
Web Ontologies
liesp.insa‐lyon.fr
•For example, assume there are only four predefined
categories, it specifies Match Rankfor multi‐class soft
ontology classification of an arbitrary instance ontology Oa
across all domain ontologies of interest.
34
A Semantic approach for Classification of
Web Ontologies
MatchRank Oa {
(book_ontology, 0.38),
(journal_ ontology, 0.58),
(ScientificMagazine_ontology, 0.24),
(ConferenceProceeding_ontology, 0.74)
}
liesp.insa‐lyon.fr
35
Example of Multi‐Class Soft Classification
Proceeding
Soft Classification
Arbitrary
Ontology
Journal
Book
Magazine
Rank: 0.24
Rank: 0.58
Rank: 0.74
Rank: 0.38
liesp.insa‐lyon.fr
•Match Rank Calculations by ONTCLASSIFIRE
–The ONTCLASSIFIREgets the arbitrary ontology Oa
for
classification. It starts the semantic similarity computation
between the Oa and domain ontologies {Od1, Od2,…,Odn} of
predefined categories.
–It employs all the syntactic, structural and semantic
knowledge present in the ontologies to compute the
match rank so that arbitrary ontology should be assigned
with a pre‐defined category label.
36
A Semantic approach for Classification of
Web Ontologies
liesp.insa‐lyon.fr
Where,
–Lcc’: Concept label similarity between c and c’
–Dcc’: Datatype properties similarity between c and c’
–Occ’: Object properties similarity between c and c’
–Pcc’: Parent concepts similarity c and c’
–Hcc’: Children concepts similarity c and c’
–Acc’: DL Axiom similarity between c and c’
37
A Semantic approach for Classification of
Web Ontologies
Sim (c, c’) = αLcc’+ βDcc’+ γOcc’+ µPcc’+ ΘHcc + ΩAcc’
liesp.insa‐lyon.fr
•Once the similarities between the concepts of domain
ontology Od
and arbitrary ontology Oa
are calculated,
ONTCLASSIFIREthen calculates the match rank between Od
and Oa
by aggregating the weights of Sim (c, c’)as shown
below.
38
A Semantic approach for Classification of
Web Ontologies
MatchRank(Od
, Oa) = ∑
i=1..n
Sim (c, c’)i
liesp.insa‐lyon.fr
•Concept Label Similarity (Lcc’).
–Label of concept is highly significant that comprise the
utmost weight in the description of concepts. Lcc’
computes the correspondences between labels of
concepts c and c’of ontologies Oa
and Od
.
•Concept Properties Similarity (Dcc’and Occ’).
–It is computed on basis of (i) Property name (ii) Range of
property and (iii) Tags associated with that property.
–Likewise, similarity between object properties (Occ’) is
computed between concepts.
39
A Semantic approach for Classification of
Web Ontologies
liesp.insa‐lyon.fr
•Concept Parent and Children similarity (Pcc’and Hcc’).
–OWL ontology
•starts from top concept Thingthat captures
everything.
•allows multiple inheritances,
•Hence, Parent similarity requires computation of
correspondences between all parent concepts.
–Pcc’analyses whether the parents of concept c and c’
are semantically similar or not
–Hcc’checks their children concept similarity.
40
A Semantic approach for Classification of
Web Ontologies
liesp.insa‐lyon.fr
•DL Axiom Similarity (Acc’).
–OWL classes are described through so‐called class
descriptions (equivalent to DL axioms), e.g.,
–DL axioms define the context of concept increases the
ability of classifier to make more accurate reasoning on
concept for their semantic similarities
41
A Semantic approach for Classification of
Web Ontologies
Publication concept can be represented as
{ Thesis ∏∃WrittenBy.Student } OR
{ Paper ∏∃ReviewedBy.Committee ∏∃>8 haslimit.Pages }
accordingly to its context.
liesp.insa‐lyon.fr
42
Outlines
Introduction
Related Work
Ontology Classification Problem
Semantic based Ontology Classifire
Preliminary Results/Experiment
Lessons to Learn and Conclusion
42
liesp.insa‐lyon.fr
•Firstly, we built the hierarchical category ontology for
Librarythat contains several categories, e.g., Book,
Proceeding, Thesis, Journal, etc.
43
Experiment –Preliminary Results
Magazine
Journal
Proceeding
Th
esis
Book
Library
liesp.insa‐lyon.fr
•Secondly, each category is elaborated with the domain ontology
that enriches the semantics and differentiates the categories
themselves. Figure 2 shows the fragment of category ontology, and
domain ontologies of two categories Bookand Proceeding.
44
Experiment –Preliminary Results
BookProceeding
category
These categories are overlapping and hence The
domain ontologies share common vocabulary in
terms of concepts (e.g., Author, Publisher, etc.),
properties (ISBN, Title, Price, etc.)and relations
(e.g., collectionOf, formatType, etc.)between
them.
These
cat
e
gories
are
overlapping
and
he
nce
The
domain
ontologies
sha
r
e
common
vocabulary
in
terms
of
concepts
(e.g.,
Author
,
Publisher
,
etc.)
,
properties
(
ISB
N
,
Title
,
Price
,
etc.)
and
relations
(e.g.,
collectionOf
,
formatType
,
etc.)
be
twee
n
them.
liesp.insa‐lyon.fr
•These categories are overlapping and hence the domain
ontologies share common vocabularyin terms of concepts
(e.g., Author, Publisher, etc.), properties
(ISBN, Title, Price,
etc.) and relations (e.g., collectionOf, formatType, etc.)
between them.
•The differentiated concepts and properties between these
categories are assigned weights in domain ontologies so
that classification can be done more accurately on basis of
specific differentiating aspects of each category.
–For example, concepts (Academic_Papers, Organizing
Committee, Conference, etc.) and properties (presentedAt, peer
ReviewedBy, feedback,
etc.)
differentiate the category
Conference Proceedingfrom Book, and hence provided with
some weights.
45
Experiment –Preliminary Results
liesp.insa‐lyon.fr
•When the arbitrary ontology Oa
comes,
ONTCLASSIFIREcomputes the similarities
between the domain ontologies and arbitrary
ontology.
•Finally, on basis of calculated highest match rank,
arbitrary ontology is assigned a label Proceeding,
but the match rank is preserved in the knowledge
base which could be used for future perspective
of query answering for ontology retrieval.
46
Experiment –Preliminary Results
liesp.insa‐lyon.fr
47
Arbitrary
ontology
Arbitrary
ontology
Experiment
–
P
reliminary
Results
liesp.insa‐lyon.fr
48
Experiment –Preliminary Results
Match Rank
Calculation
liesp.insa‐lyon.fr
49
Outlines
Introduction
Related Work
Ontology Classification Problem
Semantic based Ontology Classifire
Preliminary Results/Experiment
Lessons to Learn and Conclusion
49
liesp.insa‐lyon.fr
•Classification practice has long been adopted in digital libraries and
information systems to facilitate user in clarifying his information need
and to structure search results for browsing.
•From last decade, it has received great attention in the context of helping
users to cope with the vast amount of information on the Web. For
emerging semantic web, complex structure and semantics of ontologies
presents additional challenges as compared to traditional text
classification and web page classification.
•One of the core challenges of current semantic web research is to develop
semantic web portals that assist individual knowledge engineers to search
terms and ontologies, and serve tools and web‐agents seeking data and
knowledge in sound semantic manner
•But, state‐of‐the‐art ontology semantic web portals are not effectively
meeting the demands of ontology classification for ontology based
enterprise business applications and upcoming semantic web.
50
Conclusion
liesp.insa‐lyon.fr
•Classification of ontologies is essential for ontology, concept or
information management and retrieval tasks on semantic web.
•It improves the quality of web search for specific ontologies and
concepts.
•In addition, classification of the web ontologies is crucial tasks to promote
focused crawling for ontology retrieval and concept specific modular
ontology analysis.
•Usually search results are presented in a ranked list for assistance to
users. Soft classification mechanism exploited by ONTCLASSIFIREcould be
more useful to users in this aspect.
•The use of ontology matching for ontology classification provides higher
accuracy of classification especially in the case of overlappingontologies.
•This work can benefit construction, maintenance or expansion of
ontologies directories on the semantic web.
51
Conclusion
liesp.insa‐lyon.fr
•ONTCLASSIFIREmakes use of context specific similarity measures to fit the
ontologies into a predefined directory of general categories.
•It replaces the plain text classification algorithm in the process of ontology
classification with ontology specific classification algorithm.
•Instead of using keyword search with bag‐of‐words, it uses basic domain
ontology for each predefined category and benefit from ontology
matching research to find the correspondences between the domain
ontology and arbitrary ontology for classification purpose.
•Preliminary results show that, ONTCLASSIFIRE, forms a suitable basis for
ontology classification. One of the ongoing researches is to train the
ONTCLASSIFIREon real world ontology repository dataset such as dmoz,
and present the empirical result of our semantic based techniquefor
classification of ontologies. At the same time, we are building the retrieval
mechanisms of proposed framework.
52
Conclusion and Future Direction
liesp.insa‐lyon.fr
References
•[Bosch, 07] Bosch, A., Zisserman, A., Munoz., X.: Image classification using random forests and
ferns. In ICCV, 2007
•[Chakrabarti, 02] Chakrabarti, S., Joshi, M.M., Punera, K., Pennock, D.M.: The structure of broad
topics on the web. In Proc. 11th Intl. Conference on WWW, NY, 251–262. ACM 2002
•[Ding, 05] Ding, L., Pan, R., Finin, T., Joshi, A., Peng, Y., Kolari, P.: Finding and Ranking Knowledge on
the Semantic Web, In Proc. 4th International Semantic Web Conference. Springer LNCS 3729, 156–
170 2005
•[Dmoz, 07] Corporation, N.C., (2007). The dmoz open Directory Project (ODP).
http://www.dmoz.com/.
•[Ehrig, 05] Ehrig , M., and Maedche, A.: Ontology‐Focused Crawling of Web Documents, In Proc.
ACM symposium on Applied computing, Melbourne, Florida, 1174 –1178 2005
•[Fahad, 07] Fahad, M., Qadir, M.A., Noshairwan, M.W., Iftakhir, N.: DKP‐OM: A Semantic Based
Ontology Merger. In Proc. 3rd International Conference I‐Semantics 2007, Graz, Austria, 313‐322
•[Glover, 02] Glover, E.J., Tsioutsiouliklis, K., Lawrence, S., Pennock, D.M., Flake, G.W.: Using web
structure for classifying and describing web pages. In Proc. 11th Intl. Conference on World Wide
Web, NY, 562–569 ACM 2002
•[Grobelnik, 05] Grobelnik, M., Mladenić, D.: Simple classification into large topic ontology of Web
documents, Journal of Computing and Information Technology,Vol. 13(4) 2005
•[Mitchell, 97] Mitchell, T.M.: Machine Learning. New York, McGraw‐Hill 1997
•[Ontolingua, 10] Ontolingua Website,
http://www.ksl.stanford.edu/software/ontolingua/
53
liesp.insa‐lyon.fr
References
•[Pan, 06] Pan, J.Z., Thomas, E., Sleman, D.: ONTOSEARCH2: Searching and Querying Web
Ontologies. In: Proc. of WWW/Internet’06. pp. 211–218 2006
•[Patel, 03] Patel, C., Supekar, K., Lee, Y., Park, E.K.: OntoKhoj: A semanticWeb portal for ontology
searching, ranking and classification. In Proc. of the 5th ACM Intl. Workshop on Web Info. and Data
Management, 58–61 2003
•[Peng, 02] Peng, X., Choi, B.: Automatic web page classificationin a dynamic and hierarchical way,
In Proc. IEEE International Conference on Data Mining (ICDM’02), Washington, DC, 386–393. IEEE
CS 2002
•[Qu, 06] Qu, H., Pietra, A.L., Poon, S.: Automated blog classification: Challenges and pitfalls,
Computational Approaches to Analyzing Weblogs, AAAI Press 2006
•[Seidenberg, 06] Seidenberg, J., Rector, A.: Web ontology segmentation: Analysis, classification and
use. In Proceedings of the 15th International Conference on the World Wide Web (WWW). ACM,
New York, 13–22 2006
•[Supekar, 03] Supekar, K., Patel, C., Lee, Y.: Characterizing Quality of Knowledge on Semantic Web,
University of Missouri ‐Kansas City, Technical Report, 2003
•[Taghva, 03] Taghva, K., Borsack, J., Coombs, J., Condit, A., Lumos, S., Nartker,T.: Ontology‐based
Classification of Email, In Proc. Intl. Conference on Information Technology: Computers and
Communications,194‐200 2003
•[Wu, 03] Wu, S.H., Tsai, T.H., Hsu, W.L.: Text Categorization Using Automatically Acquired Domain
Ontology, In Proc. 6th international workshop on Information retrieval with Asian languages, 138‐
145 Japan 2003
•[Yahoo, 07] Yahoo!, Inc. (2007). Yahoo! http://www.yahoo.com/.
54
liesp.insa‐lyon.fr
55
55
A Semantic Approach for Classification
of Web Ontologies
M. Fahad, N. Moalla, A. Bouras
Thankyou
for
your
attention!
liesp.insa‐lyon.fr
56
56
A Semantic Approach for Classification
of Web Ontologies
M. Fahad, N. Moalla, A. Bouras
Thankyou
for
your
attention!
?
Enter the password to open this PDF file:
File name:
-
File size:
-
Title:
-
Author:
-
Subject:
-
Keywords:
-
Creation Date:
-
Modification Date:
-
Creator:
-
PDF Producer:
-
PDF Version:
-
Page Count:
-
Preparing document for printing…
0%
Commentaires 0
Connectez-vous pour poster un commentaire