INTEGRATING SOCIAL WEB WITH SEMANTIC WEB

Ontology Learning and Ontology Evolution from Folksonomies

Ademir Roberto Freddo and Cesar Augusto Tacla

Universidade Tecnológica Federal do Paraná, Av. Sete de Setembro 3165, Curitiba, Paraná, Brazil

Keywords: Social Web, Semantic Web, Ontology Learning, Ontology Evolution, Folksonomy, Ontology Alignment.

Abstract: In this paper, we present an approach for integrating Social Web with Semantic Web by combining the

easiness of annotation of resources in the Social Web and the expressiveness of ontologies to describe the

resources in the Semantic Web. Our approach combines ontology learning and ontology evolution

techniques to provide an integrated Web. Besides, we show how ontology alignment can be used to enrich

ontologies in this context.

1 INTRODUCTION

The Semantic Web requires resources to be

annotated with machine understandable metadata

(Berners-Lee et al, 2001) being ontologies the

knowledge representation technique most used

nowadays to describe such metadata.

Even knowledge engineers and experts have

some difficulty of maintaining consistency between

resources and the ontologies. It is necessary to catch

the changes of web resources and keep update the

ontology. Besides, the lack of an imagined class or

wrong classification when users annotate a Web

resource with applications built on ontologies is a

recurrent problem.

The knowledge acquisition bottleneck has

limited ontology use. Ontology development starts

with an initial ontology which is later revised,

refined and filled with details (Heflin et al, 1999;

Noy and McGuinness, 2001). Besides, new

information which was previously unknown or

unavailable needs to be added to the initial ontology.

Nowadays changes to ontologies have to be captured

and introduced by knowledge engineers (Zablith et

al, 2008).

On the other hand, in the Social Web, social

tagging systems such as Flickr

(http://www.flickr.com) for photo sharing, and

delicious (http://delicious.com) for social

bookmarking are becoming more popular in the

Web. The reason for their immediate success is the

fact that no specific skills are needed for annotating.

Users annotate, assign tags (any keyword, label),

and categorize web resources easily and freely

without using or even knowing taxonomies or

ontologies. In the social tagging systems, user’s

resources and associated tags constitute the

personomy. The collection of personomies

constitutes a folksonomy (Jaschke et al, 2008). The

folksonomy is dynamic as long as users learn new

things and review their personomies, including and

excluding their tags. Creating personomies is easy

and does not require expert users. However, it let

users to introduce ambiguities. Thus, in

folksonomies content retrieval activities such as

searching are limited, because results can present

low recall and precision. So, in the Social Web the

meaning of the tagging data has limited useful

reasoning with the data.

1.1 Folksonomy and Ontology

Folksonomy is created by open and uncontrolled

systems (social tagging systems) where users can

annotate resources with different tags depending on

their social or cultural backgrounds, expertise and

perception of the world (Belelman et al, 2006;

Golder and Huberman, 2005; Peterson, 2006; Wu H.

et al, 2006).

Ontology is “an explicit and formal specification

of a conceptualization” (Gruber, 1993). Ontologies

specify common conceptualizations, independent of

data model, so people can align their systems

semantically by adopting ontologies.

247

Roberto Freddo A. and Tacla C. (2009).

INTEGRATING SOCIAL WEB WITH SEMANTIC WEB - Ontology Learning and Ontology Evolution from Folksonomies.

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development, pages 247-253

DOI: 10.5220/0002303102470253

 SciTePress

In contrast to ontology, folksonomy does not

explicitly state shared conceptualizations. It is a

free-form annotation of web resources, done by

users, and without the constraints of a predefined

taxonomy or ontology (Wu H et al, 2006). Thus, in

the Social Web, the meaning of the tagging data is

completely unspecified. We need technology for

reasoning with the folksonomies in such a way that

computations can discover and conclude new things

(Gruber, 2008).

The integrated Web combines the facilities in

annotating web resources with tags that

characterizes Social Web applications with

ontologies to better describe the resources in the

Semantic Web. According to Tim Berners-Lee

(Berners-Lee et al, 2001) "the Semantic Web is not a

separate Web but an extension of the current one, in

which information is given well-defined meaning,

better enabling computers and people to work in

cooperation”.

1.2 Contribution and Organization of

the Paper

Our approach aims at extracting structured data

(ontologies) from unstructured data (folksonomies).

As result of our approach, we obtain an ontology

whose elements are linked to their source tags

enabling to trace folksonomy changes back to the

ontology.

The folksonomies in Social Web can provide

new impulse to Semantic Web in by reducing the

burden on users and engineers in tasks related to

knowledge engineering (e.g. knowledge acquisition,

Web resource annotation, and ontology

construction). On the other hand, the Semantic Web

can improve inferences and provide better query

results in Social Web.

The remainder of the paper is organized as

follows. Section 2 describes the basic definitions and

current developments in ontology learning and

ontology evolution from folksonomies. In section 3,

we describe the approach to ontology learning and

evolution. Section 4 presents a case study. Finally,

Section 5 gives a conclusion and presents envisaged

works.

2 STATE OF THE ART

In this section we describe the state of the art in

ontology learning and evolution.

2.1 Ontology Learning

According to Maedche and Staab (2001), “ontology

learning greatly facility the construction of

ontologies by the ontology engineer. It is a task of

(semi)-automatically construct an ontology by using

machine learning or data mining algorithms that are

applied on data”.

In folksonomies, the use of tags by people with

common interests tends to converge to a shared

vocabulary. Following this intuition, a variety of

approaches have been proposed to discover shared

conceptualizations that are hidden in a folksonomy.

Some of them (Schmitz, 2006; Wu et al., 2006;

Belelman et al., 2006; Mika, 2005) analysis the co-

occurrence of tags. Schmitz (2006) finds candidate

subsumption relations, Wu et al (2006) and

Belelman et al. (2006) create clusters of tags, and

Mika (2005) builds graphs relating tags. However

these cited approaches focus on finding groups of

related tags rather than identifying the semantics of

those relations.

Another set of recent approaches (Angeletou et

al, 2007; Basso and da Silva, 2008; Specia and

Mota, 2007) build ontologies from folksonomies

thus going beyond the mentioned approaches that

identify implicitly inter-related tags. Angeleton et al

(2006) propose a method to enrich the tag space of

folksonomies by exploring ontologies. Basso and da

Silva (2008) presents a proposal for the ontology

construction/evolution from the folksonomies based

on WordNet. Specia and Mota (2007) propose the

integration of folksonomies and ontologies to enrich

tag semantics – identify semantic relation between

tags using ontologies available on the semantic web.

These approaches has some limitations:

difficulty in finding online ontologies; syntactic

mapping to link tags to ontology concepts; different

ontologies reflect different views which often lead to

contradictory; limited reasoning because of plain

structure of folksonomies; and they not focus on the

ontology evolution.

In social tagging systems (e.g.

http://www.flickr.com, http://delicious.com) there is

no formal semantic and no formal agreement on the

representation of the tagging process. This means

that every system uses a different format to publish

its tagging data.

Thus, other approaches (Knerr, 2006; Gruber,

2007, Kim et al, 2007) propose a solution for tag

data representation. Gruber’s conceptual model

(Gruber, 2007) describes tagging as a relation

between an object (resource – bookmark, picture), a

tag, a tagger (a person or agent that created the link

KEOD 2009 - International Conference on Knowledge Engineering and Ontology Development

248

between the tag and the object) and a source (the

space where the tagging action has been performed –

flickr, delicious). Knerr (2006) develop an ontology

using the tripartite tagging (user, resource, tag)

model Newman et al, (2005). The Social Semantic

Cloud of Tags (SCOT) (Kim et al, 2007) ontology

aims to describe the structure and the semantics of

tagging data and to offer social interoperability of

the data among heterogeneous sources.

2.2 Ontology Evolution

Ontology evolution is defined by (Haase and

Stojanovic, 2005) as the “timely adaptation of an

ontology to the arisen changes and the consistent

management of these changes”. Ontology evolution

is a process that supports the enrichment of the

ontology by adding new entities (concepts,

properties, and instances) or by modifying existing

entities when new knowledge is acquired.

Ontology changes may come from explicit and

implicit requirements (Cimiano and Volker, 2005).

Explicit requirements are generated by ontology

engineers. Implicit requirements are reflected in the

behavior of the system and can be induced by

applying change discovery methods from existing

data (Cimiano and Volker, 2005; Stojanovick,

2004). Stojanovick (2004) defines three types of

change discovery: structure-driven deduces changes

from the ontology structure itself, usage-driven

identifies changes from the usage patterns creates

over a period of time, and data-driven generates

changes by modifications to the underlying data

(text documents, database) that represents the

knowledge modeled by an ontology.

The current state of the art in ontology evolution,

as well as a list of existing tools that aid the process

can be found in (Haase and Sure, 2004). Some of

these approaches are simple ontology editors, like

Protégé (Noy et al, 2000) and OilEd (Bechhofer et

al, 2001). One set of approaches (Alani et al, 2006;

Bloehdorn et al, 2006; Novacek et al, 2007;

Novacek et al, 2008, Ottens and Glize, 2007;

Cimiano and Volker, 2005; Zablith, 2007, Zablith et

al, 2008) in ontology evolution identify potential

novel information that should be added to the

ontology by exploiting the changes occurring in the

various data sources. Such approaches do not

consider folksonomies as an information source

where ontology changes can be discovered. In this

work, this is exactly the case, as we focus on change

capturing from data-driven implicit requirements.

Our collaborative approach for ontology evolution

starts with an existing ontology (base or domain

ontology) and supports the acquisition of new

knowledge from folksonomies when users change

their tags as long as their vocabularies change. We

evolve ontologies exploiting folksonomy versioning

and linking learned ontology entities to the source

tags in the folksonomy.

3 SOCIAL WEB WITH

SEMANTIC WEB

This section presents the approach to ontology

learning and evolution from folksonomies.

3.1 Ontology Learning from

Folksonomies

The ontology learning process from folksonomies is

articulated in the following phases (Figure 1):

populating the tag ontology, identifying relations

between tags, and interacting with user.

Figure 1: Ontology learning from folksonomies.

The populating tag ontology phase consists of

building a representation for folksonomy tags based

on entities defined in SCOT, an ontology to describe

the structure and the semantics of folksonomies.

The SCOT ontology uses concepts and

proprerties of Newman’s model (Newman et al,

2005).

After populating SCOT ontology, we identify for

each pair of tags, ((tag

,tag

), (tag

,tag

), (tag

,tag

(tag

,tag

), (tag

,tag

), (tag

,tag

)), the tag type

(whether concept or instance), and the relation

between them based on properties defined in SCOT

such as textual description, synonym, and spelling

INTEGRATING SOCIAL WEB WITH SEMANTIC WEB - Ontology Learning and Ontology Evolution from

Folksonomies

249

variant. In order to identify tag types, we use

Text2Onto (Cimiano and Volker, 2005). For tag

relations, we use WordNet (Miller, 1990) for

identifying meronyms, hyponyms, synonyms, and

hyperonyms, and the Hwang’s work (2007) for

performing analysis of classes’ hierarchy.

A tag can be described as an instance or a

concept in the base ontology. Thus, it is possible to

obtain the following types of pairs: (instance,

instance), (instance, concept), and (concept,

concept). In pairs of type (instance, concept), we

verify if the tag instance can be considered an

instance of the concept tag. For example, for the pair

of tags (Porto, University), university is the concept

and Porto is an instance (University of Porto). So,

we add to the base ontology the concept University

and the instance Porto. For pairs (concept, concept),

we identify meronym, hyponym, and hyperonym

relations. Finally, for pairs (instance, instance), we

identify the concepts related to the instances. In this

case we create a new concept or use an already

identified concept. The approach to ontology

learning from folksonomies is semi-automatic. It

suggests to the user concepts, instances, properties,

and relations between concepts. In the interacting

with user phase, the user takes all decisions

concerning the creation of concepts, instantiation of

concepts and relations. After the user decision

taking, the base ontology is created. Each entity in

the base ontology is linked by means of the

reference metaproperty to the source tag in order to

maintain the traceability with the folksonomy.

3.2 Ontology Evolution from

Folksonomies

Figure 2 shows our approach to evolve ontologies

from folksonomies. A folksonomy is produced by

user tagging activity in any social tagging system.

Tags represent the domain according to the users’

perspective and they are formally described using

SCOT ontology. The task ontology contains the

rules to extract data from folksonomies and to

populate the SCOT. The base ontology is created by

ontology engineers or some ontology learning

process. Our purpose is to respond to changes in the

folksonomies to update the base ontology. If the

folksonomy is changed, the base ontology may also

be modified. Our approach starts with reading a

folksonomy in order to extract new and relevant

information to be added to or removed from the base

ontology. As the SCOT is populated with

information from the folksonomy, the addition or

removal of instances means possible changes to be

done in the base ontology.

Figure 2: Approach to ontology evolution.

3.2.1 Add Concepts in Base Ontology

When new tags are added to the folksonomy, the

base ontology may have to be updated with new

entities. After identifying the changes, we identify

similarity relations between the extracted tags and

the entities in the base ontology (Figure 3) using an

ontology alignment method. With this method, we

know which parts of the base ontology are affected

by the changes in the folksonomy.

Besides, we identify the proper position where

the new entity should be added. In this work we use

the partial ontology alignment method named

Partial Ontology Alignment Method - POAM

(Freddo et al., 2007). For example, a new instance

and its associated properties in the SCOT ontology

are compared to the other SCOT’s instances

obtaining a set of similar instances. We search in the

base ontology the concepts linked to the instances in

the set by the reference metaproperty. At this

moment, the user has to decide which relation the

new instance has with the linked concepts of the

base ontology: none, subsumption, equivalence,

sibling or instance of.

3.2.2 Remove Concepts in Base Ontology

Each concept in the base ontology is referenced by

one or more instances of SCOT’s concepts. When a

tag is removed from the folksonomy, we verify

which concepts are linked to this tag. If the tag is

linked only to one concept in the base ontology, we

suggest to the user to remove the concept from the

base ontology. If the tag is linked to two or more

concepts in the base ontology, we do not remove the

KEOD 2009 - International Conference on Knowledge Engineering and Ontology Development

250

concept. We only update the reference meta-

property. As in the ontology learning, this approach

to ontology evolution creates suggestions for

changes which may be help knowledge engineers in

taking final decisions.

Figure 3: Ontology evolution and ontology alignment.

4 CASE STUDY

A case study is presented to demonstrate the

complete approach comprising ontology learning

and evolution.

4.1 Case Study to Ontology Learning

from Folksonomies

We demonstrate the ontology learning with the tags

hotel, accommodation, room, luxury, Paris were

retrieved from some folksonomy in the tourism

domain. We use these tags to populate the SCOT

ontology. Figure 4 shows the SCOT ontology where

each instance of the tag concept has some properties

such as reference_tag, equivalentTag,

spelling_variant, used_by, and associatedTag_in.

The reference_tag property maintains the

traceability between a tag in the folksonomy and the

instance of the tag in SCOT.

After populating the SCOT ontology, we identify

the relation between pair of tags. In this example, we

have the following pairs: (hotel, accommodation),

(hotel, room), (hotel, luxury), (hotel, Paris),

(accommodation, room), (accommodation, luxury),

(accommodation, Paris), (room, luxury), (room,

Paris), and (luxury, Paris).

Figure 4: Instances of the SCOT’s tag concept.

Figure 5: Initial ontology.

Figure 5 shows the initial ontology with some

relations such as instance_of, has, and is a. Each tag

can be an instance or concept in the initial ontology.

With the suggestions represented in the initial

ontology (Figure 5), the user builds the base

ontology. In the base ontology, each concept has

properties and meta-properties (properties in SCOT

are meta-properties for concepts in the base ontology

and temporary concept). For instance, the concept

hotel has properties hasCity, hasRoom, and

hasClassification. The same concept has meta-

properties reference_tag, equivalentTag,

spelling_variant, etc.

4.2 Case Study to Ontology Evolution

from Folksonomies

We demonstrate the ontology evolution approach

with the inclusion of new tags (hostel, address) in

folksonomy (Figure 6). The new tags are instances

of SCOT’s tag concept. Based on the properties

defined for the tag concept in the SCOT ontology,

each tag can be seen as a temporary concept with

meta-properties. Thus, we align hostel and address

with concepts in the base ontology based on the

INTEGRATING SOCIAL WEB WITH SEMANTIC WEB - Ontology Learning and Ontology Evolution from

Folksonomies

251

POAM’s similarity value computed with several

metrics including property similarity ones. The

POAM detects the following alignments: hostel and

hotel, and address and city (Figure 7).

tag ontology

SCOT

Instances of

tag concepts

equivalentTag

hostel

address

spelling_variant

used_by

associatedTag_in

reference_tag

…...

properties

Populate tag

ontology (SCOT)

folksonomy

hotelaccommodation

room luxury Paris

hostel address

Base

ontology

accommodation

hotel

room

luxury

Paris

city

hasClassification

hasCity

is a

instance_of

hasRoom

hostel

address

Ontology

Alignment

temporary

concepts

meta-properties

Figure 6: Evolving an ontology from new tags in the

folksonomy.

Figure 7: Similarity between temporary concepts and

concepts in the base ontology.

Hostel and address are temporary concepts that

were aligned with concepts city, accommodation,

room, hotel and luxury in base ontology. Based on

the found alignments, the user knows the proper

position where the concepts hostel and address are

added. Then we identify the relations between

following pair of tags: (hostel, hotel), (hostel,

accommodation), (address, city).

5 CONCLUSIONS

In Social Web, we have users building their

personomies online. However, the meaning of tags is

completely unspecified. Ontologies can describe

semantically such data.

By combining the facilities in annotating Web

resources in the Social Web and the expressiveness

power of Ontologies to describe resources in the

Semantic Web, we could provide an integrated Web.

In this work we describe an approach for

combine the Social Web and the Semantic Web.

According to Gruber (2008), “the challenge for the

next generation of the Social and Semantic Web is to

find the right match between what is put online and

methods for doing useful reasoning with the data”.

Motivated by the challenges of ontology

engineering and inspired by the success of social

web applications, we presented an approach to

ontology learning and evolution from folksonomies.

We use ontology alignment to support ontology

enrichment (add or remove entities) when changes

are detected in the folksonomy.

The implementation of this approach is currently

in progress. In a near future, we intend to evaluate it

analyzing the inferences that become possible with

the integration, the precision and recall in queries,

and the degree of user overloading with tasks related

to knowledge engineering related to the ontology

development and evolution.

ACKNOWLEDGEMENTS

The authors thankfully acknowledge the financial

support granted for this study by Fundação

Araucária (AMO 9902 – conv. 234/2007).

REFERENCES

Alani, H., Harris, S., O'Neil B. 2006. Winnowing

ontologies based on application use. In Proceedings of

3rd European Semantic Web Conference (ESWC),

Montenegro.

Angeletou, S., Sabou, M., Specia L., Mota, E. 2007.

Bridging the Gap Between Folksonomies and the

Semantic Web: An Experience Report. In: Workshop

Bridging the Gap between Semantic Web and Web 2.0

- 4th European Semantic Web Conference, Insbruck.

Workshop Bridging the Gap between Semantic Web

and Web 2.0, p. 30-43.

Basso, C. A. M., da SILVA, S. R. P. 2008. Uma Proposta

para a Evolução de Ontologias a partir de

Folksonomias. In: Webmedia - Workshop de Teses e

Dissertações, 2008, Vila Velha - ES. Webmedia.

Sociedade barsileira de Computação, v. 1. p. 197-200.

Belelman, G., Keller, P., Smadja F., 2006. Automated Tag

Clustering: Improving search and exploration in the

tag space. In Proc. of the Collaborative Web Tagging

Workshop.

KEOD 2009 - International Conference on Knowledge Engineering and Ontology Development

252

Berners-Lee, T., James H., Ora L., 2001. The Semantic

Web. Scientific American Magazine.

http://www.sciam.com/article.cfm?id=the-semantic-

web&print=true. Retrieved on 2008-03-26.

Bloehdorn, S., Haase, P., Sure, Y., Voelker J. 2006.

Ontology Evolution. In: John Davies, Rudi Studer and

Paul Warren, Semantic Web Technologies - Trends

and Research in Ontologybased Systems. pp. 51-70.

John Wiley & Sons.

Cimiano, P., Volker, J., 2005. Text2Onto - a Framework

for Ontology Learning and Data-driven Change

Discovery. In:Proc. NLDB, LNCS 3513, pp. 227-238,

Springer, Alicante.

Freddo, A. R., Brito, R. C., Gimenez-Lugo G., Tacla, C.

A., 2007. Partial and Dynamic Ontology Mapping

Model in Dialogs of Agents. EPIA Workshops: 347-

356. Services. AI Magazine 26(1): 71-82.

Golder S., Huberman B.A., 2005. The Structure of

Collaborative Tagging Systems. HPL Technical

Report.

Gruber, T. R., 1993. Toward Principles for the Design of

Ontologies Used for Knowledge Sharing., In Formal

Ontology in Conceptual Analysis and Knowledge

Representation, Nicola Guarino et Roberto Poli (eds.),

Kluwer Academic Publishers, International Workshop

on Formal Ontology, Padova, Italy.

Gruber T. R., 2007. Ontology of folksonomy: a mash-up

of apples and oranges. Intl Journal on Semantic Web

& Informatin Systems.

Gruber, T. R., 2008. Collective knowledge systems:

Where the Social Web meets the Semantic Web. Web

Semantics: Science, Services and Agents on the World

Wide Web. Volume 6, Issue 1, Semantic Web and Web

2.0, p. 4-13.

Haase, P., Sure, Y., 2004. D3.1.1.b State of the Art on

Ontology Evolution Available at: http://www.aifb.uni-

karlsruhe.de/WBS/ysu/publications/SEKT-

D3.1.1.b.pdf

Haase, P., Stojanovic, L. 2005. Consistent Evolution of

OWL Ontologies. In: Proceedings of the 2nd

European Semantic Web Conference. Heraklion,

Greece.

Heflin, J., Hendler, J., Luke, S., 1999. Coping with

Changing Ontologies in a Distributed Environment. In

Proceedings of the Workshop on Ontology

Management of the 16 National Conference on

Artificial Intelligence, pp. 74—79.

Hwang Suk-Hyung. 2007. A Triadic Approach of

Hierarchical Classes Analysis on Folksonomy Mining.

Intl. Journal of Computer Science and Network

Security (IJCSNS).

Jaschke R., Hotho, A., Schmitz C., Ganter, B., Stumme

G., 2008. Discovering shared conceptualizations in

folksonomies. In Web Semantics: Science, Services

and Agents on the World Wide Web. vol. 6, Issue 1,

Semantic Web and Web 2.0, p. 38-53.

Kim H. L., Yang S. K., Breslin, J. G., Kim H. G., 2007.

Simple algorithms for representing tag frequencies in

the scot exporter. In IAT, pages 536-539. IEEE

Computer Society.

Knerr, T., 2006. Tagging Ontology- Towards a Common

Ontology for Folksonomies. Available at:

http://tagont.googlecode.com/files/TagOntPaper.pdf.

Maedche, A., Staab, S. 2001. Ontology learning for the

semantic web. IEEE Intell. Syst. 16 (2) 72–79.

Mika, P., 2005. Ontologies are us: A unified model of

social networks and semantics. Proceedings of the 4th

International Semantic Web Conference (ISWC),

pp.522--536.

Miller, G.A. 1990. WordNet: An on-line lexical database.

In International Journal of Lexicography 3(4), 235–

312.

Newman R., Ayers, D., Russell. S. 2005. Tag ontology,

Available at:

http://www.holygoat.co.uk/owl/redwood/0.1/tags/.

Novacek, V., Laera, L., Handschuh, S. 2007. Semi-

automatic integration of learned ontologies into a

collaborative framework. In International Workshop

on Ontology Dynamics (IWOD).

Novacek, V., Laera, L., Handschuh, S., Zemanek, J.,

Volkel, M., Bendaoud, R., Hacene, M. R., Toussaint,

Y., Delecroix, B., Napoli. A., 2008. D2. 3.8 v2 report

and prototype of dynamics in the ontology lifecycle.

Noy, N. F., McGuinness, D., 2001. Ontology development

101: a guide to creating your first ontology, Stanford

Knowledge Systems Laboratory Technical Report

KSL-01-05 and Stanford Medical Informatics

Technical Report SMI-2001-0880.

Ottens K., Glize. P., 2007. A multi-agent system for

building dynamic ontologies. In Proceedings of

AAMAS.

Peterson E., 2006. Beneath the Metadata: Some

Philosophical Problems with Folksonomy. D-Lib

Magazine, 12(11).

Schmitz, P., 2006. Inducing ontology from Flickr tags. In

Collaborative Web Tagging Workshop, 15th WWW

Conference, Edinburgh.

Specia, L., Motta, E., 2007. Integrating Folksonomies with

the Semantic Web. In: European Semantic Web

Conference (ESWC 2007), The Semantic Web:

Research and Applications. LNCS 4519, pp. 503—

517, Springer. Heidelberg.

Stojanovic, L., 2004. Methods and Tools for Ontology

Evolution. Phd thesis, University of Karlsruhe.

Wu, X., Zhang, L., Yu, Y., 2006. Exploring social

annotations for the semantic web. In 15th WWW

Conference, Edinburgh.

Wu H., Zubair, M., Maly. K., 2006. Harvesting Social

Knowledge from Folksonomies. In Proc. of

HYPERTEXT.

Zablith F., 2007. Dynamic ontology evolution. In:

International Semantic Web Conference (ISWC)

Doctoral Consortium, Karlsruhe, Germany.

Zablith, F., Sabou, M., d'Aquin, M., Motta, E., 2008.

Using Background Knowledge for Ontology

Evolution. Workshop: International Workshop on

Ontology Dynamics (IWOD). The International

Semantic Web Conference (ISWC), Karlsruhe,

Germany.

INTEGRATING SOCIAL WEB WITH SEMANTIC WEB - Ontology Learning and Ontology Evolution from

Folksonomies

253