ONTOP

A Process to Support Ontology Conceptualization

Elis Montoro Hernandes, Deysiane Sande and Sandra Fabbri

Computing Department, Federal University of São Carlos, São Carlos, Brazil

Keywords: Ontology Engineering, Collaborative Glossary, Conceptualization Phase, Information Visualization.

Abstract: Although there are tools that support the ontology construction, such tools do not necessarily take heed to

the conceptualization phase in its need of execution resources. The objective of this paper is to present the

ONTOP Process (ONTOlogy conceptualization Process) as an effective means of enhancing the

conceptualization phase of the ontology construction. This process is supported by the ONTOP-Tool which

provides an iterative way to defining a collaborative glossary and uses a visual metaphor to facilitate the

identification of the ontology components. Once the components are defined, it is possible to generate an

OWL file that can be used as an input to other ontology editors. The paper also presents an application of

the both process and the tool, which emphasizes the contributions of this proposal.

1 INTRODUCTION

Sharing large volume of data and information

through web technologies is, currently, a constant

need. Appropriate mechanisms for this are the target

of many researches mainly in the context of

semantic web (Daconta, Obrst and Smith, 2003).

Ontologies have been a common help resource.

According to Gruber (1993), ontology is a formal

and explicit specification of the description of

concepts in a domain. Ontologies represent the

semantic of a domain and can be used by many

applications.

Based on the literature (Gruber, 1993; Gómez-

Pérez, Fernández-López and Corcho, 2004), it is

possible to identify some advantages provided by an

ontology: i) improvement of the communication

among the involved people since it leads to a

particular sense of the vocabulary and meaning of

the domain terms; ii) formalization of the knowledge

avoiding ambiguities and inconsistencies; iii)

representation of the domain knowledge allowing its

dissemination and reuse. In addition, ontology

allows the knowledge improvement making it

possible for different teams develop applications in

different moments and with different purposes.

Due to these reasons it is important that domain

experts participate in development of the ontology

process aiming to avoid mistaken definitions.

In literature there are several approaches to

ontology development (Gómez-Pérez et al., 2004).

Corcho, Fernández-López and Gómez-Pérez (2003)

discuss ontology methods that had been used since

the 90’s and comment that none of them have

reached the maturity. However, Methontology

(Fernández-López, Gómez-Pérez, Pazos-Sierra and

Pazos-Sierra, 1999) is a method that has been

considered one of the most complete (Corcho et al.,

2003).

Among the development activities, Methontology

is composed by the phases of Figure 1, which are

present in the majority of the ontology development

processes. Aiming at supporting the execution of

these phases some tools and languages were proposed

in literature. In this research we particularly use the

ontology editor Protégé-2000 (Noy, Fergerson and

Musen, 2000; “Protégé-2000”, 2010) and the

language OWL (Ontology Web Language) (“OWL”,

2009) due to the features they provide for our work

and to their acceptance in ontology area.

Figure 1: Ontology Development Activities of

Methontology. Adapted (Corcho et al., 2003).

Montoro Hernandes E., Sande D. and Fabbri S. (2010).

ONTOP - A Process to Support Ontology Conceptualization.

In Proceedings of the 12th International Conference on Enterprise Information Systems - Databases and Information Systems Integration, pages 58-65

DOI: 10.5220/0002908100580065

 SciTePress

In spite of such proposals, none of the tools

support the conceptualization phase highlighted in

Figure 1. The objective of this phase is to organize

the non structured knowledge, which was acquired

in the previous specification phase. The

conceptualization phase converts domain

information into a semi-formal specification using a

set of intermediate representations. This phase is

considered the most important phase for the

ontology identification an it initial activity is the

construction of a glossary (Corcho et al., 2003).

Apart from the glossary, many initial definitions and

ontology components, like classes and their

relationships, are established in this phase.

Considering this context, the objective of this

paper is to present the process ONTOP (ONTOlogy

conceptualization Process) to support the

conceptualization phase of the ontology definition

and ONTOP-Tool that supports the execution of this

process. ONTOP considers the use of a collaborative

glossary tool and the use of visualization, which

helps a great deal in the identification of classes. In

this proposal we are using the glossary available in

the free Moodle environment (Modular Object-

Oriented Dynamic Learning Environment)

(“Moodle”, 2009). The activities proposed in this

process are supported by ONTOP-Tool that is

responsible for the interaction between the glossary

and the visualization, allowing an easier

identification of the ontology components. The

intention of the example presented in this paper is to

explain the process and the functionalities of the tool

showing the contribution of the proposal and not

exploring the ontology properly.

The remaining of the paper is organized as

follows: Section 2 comments the importance of a

glossary for ontology definition and mentions the

main characteristics a glossary should have; Section

3 provides a brief view on visualization; Section 4

presents the ONTOP process and explains the

activities that compose it; Section 5 provides an

example of the process application supported by the

tool and Section 6 presents the conclusions and

future works.

2 THE IMPORTANCE OF A

GLOSSSARY FOR ONTOLOGY

DEFINITION

Regarding target domain, some authors indicate the

glossary as the artifact for knowledge acquisition

and documentation (Fernández-López et.al, 1999)

(Falbo, Menezes and Rocha, 1998).

A common problem associated with ontology

construction is that those who need the ontology are

specialists on the domain and often they are

geographically distributed. If that is the case, the

support of tools becomes essential in order to allow

a colaborative development and documentation of

the ontology construction. The glossary could be a

richer source of information if a greater number of

specialists participated in its production. In such

case, the glossary would synthetize concepts from

different contributors.

Another important reason for using glossaries is

that its use is much more friendly to the specialists

rather than the use of the notation applied to describe

ontology.

Concerning the main characteristics, a glossary

should satisfy the auto-reference principle which

says that terms used to describe another terms

should also be an entry of the glossary. In addition, a

glossary should satisfy the principle of minimum

vocabulary, i.e., the vocabulary should be as small

as possible and does not have ambiguities.

Falbo et al. (1998) emphasize that glossaries

should use the concept of hypertext aiming at

facilitating the navigation in the document. This

characteristic is native in the Moodle environment

and was a reason for choosing this environment in

this proposal. In addition, Moodle provides other

resources like a glossary administrator, different

permissions, discussion forums, etc., that facilitate

the collaborative work and its management.

3 INFORMATION

VISUALIZATION

Visualization is a process that transforms data,

information and knowledge in a visual form that

explores the natural visual capacity of human beings,

providing an interface between two powerful systems

of information treatment: the human brain and the

computer (Gershon, Eick and Card, 1998). Effective

visual interfaces provide a quick interaction with large

volume of data, making the identification of

characteristics, patterns and tendencies that were

masked easier.

In the literature, there are some visualization

techniques and tools that implement them (Gershon,

Eick and Card, 1998). They present advantages and

limitations according to the format and type of the

data that will be visualized and to the exploration

needs.

ONTOP - A Process to Support Ontology Conceptualization

Figure 2: ONTOP process.

In this research we use the Tree-Map technique

(Johnson and Shneiderman, 1991) that is illustrated in

Figure 2. This technique represents the data as nested

rectangles in accordance with their hierarchy. The

size of the rectangles is proportional to the number of

items that compose the next level of the hierarchy.

The size and color variation of Tree-Map makes

evident the characteristics of each set of data. This

fact enhances the visualization of large sets of data

like, for example, a glossary that contains many

terms. In addition, Tree-Map uses all the screen space

that allows the representation of a great amount of

data.

Figure 2 shows an example of Tree-Map. In this

case the technique is used in the NewsMap site to

group news in accordance with the subject

(“NewsMap”, 2009). This site uses visualization to

show world news allowing that users interact with the

site filtering the news by theme, date and country.

Figure 3: Newsmap site (“NewsMap”, 2009).

Visualization allows a broad view of the data as

well as the abstraction of new information in a

quicker way than if the analysis was done manually.

Even when the set of data is small, an appropriated

visualization allows an immediate identification of

tenuous differences in the data. Many advantages of

visualization uses can be viewed through the

researches of many authors (Chen, Kuljis and Paul,

2001) (Auvil, Llorà, Searsmith and Searsmith, 2007)

(Ichise, Satoh and Numao, 2008).

In this research the Tree-Map technique is used

to represent the terms of the glossary such that each

box represents a term and the size and the color of

the boxes represent the frequency that the term is

used in the glossary. In this case, the visualization

allows a quick identification of the most cited terms,

which are candidate to classes of the ontology.

Although there is, for example, the TreeMap

(“TreeMap”, 2009) and other tools that implement

this technique, they did not have essential resources to

help in our problem. Aiming to refine the glossary,

we needed that the visualization tool provided two

basic operations: string search and edition. This fact

leads us to implement ONTOP-Tool that will be

presented in Section 5.

4 THE ONTOP PROCESS

ONTOP is a process supported by the ONTOP-Tool

which enhances the ontology conceptualization by

making use of glossary and visualization (Figure 3).

The glossary can be constructed in a collaborative

way among the ontology stakeholders, including the

domain specialists, through an iterative process of

refinement. As the collaborative work is

fundamental, we decided to use the glossary of

Moodle environment for the reason that it provides

some management facilities. Plus the fact that it is

possible to export the terms to an XML file so that

they can be loaded in the ONTOP-Tool and also

visualized through the Tree-Map (Johnson and

Shneiderman, 1991), allowing the interaction

between the tool and the Moodle. After that, by

means of visualization information, it is possible to

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

define the initial ontology components. These

components can be exported to an OWL file

(“OWL”, 2009) and be used by a ontology editor

like Protégé-2000 (Noy et al., 2000; “Protégé-2000”,

2010) to go ahead with the ontology formalization.

The steps that compose the ONTOP process are

the following:

Step 1 – Refine Glossary: the objective of this step is

to create, refine and validate the glossary iteratively,

counting on the domain specialists’ involvement.

This is an iterative step where the glossary is

imported from and exported to the Moodle

environment as well as to and from the ONTOP-

Tool, until the glossary is finally able to represent

the domain. During this iteration the ontology

stakeholders can insert, remove or define the terms.

The Step 1 is composed by the following

activities:

1) Create a glossary in the Moodle environment;

2) Share the glossary with the ontology

stakeholders so that some specialists can

participate in the glossary definition;

3) Refine the glossary with the following actions:

(i) Export the glossary from the Moodle to an

XML file;

(ii) Import the XML file to the ONTOP-Tool

so it can be visualized by means of the

Tree-Map;

(iii) Export the glossary from the ONTOP-Tool

to an XML file;

(iv) Import the XML file to the Moodle

(v) Go back to (i) until the ontology

stakeholders come to an agreement.

Step 2 – Define ontology components: the objective

of this step is to identify, among glossary terms, the

possible ontology components and, then, classify

them in terms of class, class instance, relationship or

synonyms. At this point the contribution of the

ONTOP-Tool is to make evident the most used

terms, pinpointing then as possible ontology

components candidates.

Step 3 – Define class hierarchy: the objective of this

step is to define the hierarchy of the components

identified in step 2. The hierarchy is easily

established by the ONTOP-Tool through a drag-and

drop action.

Step 4 – Define class relationships: the objective of

this step is to attribute the relationships among the

classes. Some of these relationships are predefined

and obtained from the information generated in Step

2 and some others can be inserted by the user when

necessary.

Step 5 – Generate OWL file: the objective of this

step is to generate the OWL file which is composed

of all the information defined by the user in the

previous four steps. This file can be imported by an

ontology editor as Protégé-2000, which is used in

this research.

5 AN EXAMPLE OF USING

ONTOP AND ONTOP-TOOL

In this section we present an example of the process

application, detailing the functionalities provided by

the ONTOP-Tool. The process is illustrated based

on Experimental Software Engineering (ESE)

domain. The glossary constructed for this domain

counted on the collaboration of domain specialists

that composed the program committee of 2006’s

Experimental Software Engineering Latin American

Workshop (Fabbri, Travassos; Maldonado;

Mendonça Neto and Oliveira, 2006).

Experimental Software Engineering is a growing

area in software engineering and deals with different

types of experimental studies, for example, surveys,

case study, controlled experiment (Wohlin et al.,

2000). Due to limited space we cannot give a deep

overview of the domain. In spite this limitation, our

main objective is to explain the process steps

showing how they work and how they help the

ontology conceptualization phase.

The ESE glossary was constructed in the Moodle

environment aiming to facilitate the communication

among the program committee, which was

geographically distributed. Based on this ESE

glossary version, ONTOP was applied as illustrated

below.

Figure 4 shows the initial screen of the ONTOP-

Tool which has buttons for the functionalities

needed to execute the process steps.

To execute Step 1, for refining the glossary, the

user should use the first three buttons. Clicking on

the button “Import Moodle Glossary” the tool

uploads the XML file that contains the ESE

Glossary.

After that, the user should click on the button

“Analyse the Glossary” to visualize it like in Figure

5 where:

 each box corresponds to a term;

 each box is colored according to the term

frequency;

 clicking on a box it is possible to insert or edit

the term definition;

ONTOP - A Process to Support Ontology Conceptualization

 boxes that have a fading color represent terms

that were edited in the current visualization;

 terms can be inserted or excluded as in the

Moodle glossary.

Figure 4: ONTOP-Tool initial screen.

In this example, as the ESE glossary was already

constructed by the domain specialists, it was not

necessary much iteration to execute the refinement.

However, just to exemplify the contribution of

visualization, note that at the left top corner of

Figure 5 there are a set of boxes that are grouped

because they correspond to terms that do not have a

definition. This situation is easily identified in the

visual metaphor. To obtain this information in the

Moodle environment, the user should verify the

terms, one by one. Missing or equal definitions are

quickly identified by means of the ONTOP-Tool.

Considering the previous situation, if the user

decides to insert a definition to these terms, the color

of their boxes is faded (see Figure 6). This is an

interesting artifice of the ONTOP-Tool since every

time a color is faded in the visual metaphor it means

that the corresponding term was edited. The color

will persist faded while the user stays in the same

functionality.

If the user wishes to share the editions among the

ontology stakeholders, he should export this version

of the glossary clicking on the button “Export

Moodle Glossary” and import it again to the Moodle

environment, by means of Moodle functionality.

Figure 5: Initial glossary visualization.

Figure 6: Fade color represents edited terms.

All these activities should be repeated until the

ontology stakeholders reach an agreement. Once the

glossary is finished, the user can execute the Step 2,

clicking on the button “Identify components”, for

classifying the glossary terms as ontology

components. Figure 7 shows the screen of this

functionality, where the region to insert the

definitions is highlighted. The visualization is the

same of the previous functionality and, at this

moment, one of the contributions of visualization is

related to the size or the color of the boxes, since

they represent the frequency associated to each term.

Terms that have high frequency are candidates to

become classes of the ontology.

For example, in Figure 7, the terms Experiment,

Simulation and Survey that are highlighted in the

figure, correspond to the most referenced in the ESE

glossary; they have the largest boxes and colors that

correspond to high levels of frequency.

In fact, for the ESE domain, the term

“Experiment” is used for defining or expressing

many other terms like Controlled Experiment,

Experiment Design, Replicated Experiment, etc. The

same happens with the term “Simulation” that is

used to compose Continuous Simulation and

Dynamic Simulation, in addition to define other

terms.

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

Figure 7: Screen for defining components.

Another contribution of visualization for

defining the ontology components is related to the

search resource. In this case, when the user classifies

a term, the ONTOP-Tool uses this term as a key

word for searching other terms that use the defined

term in some way. As it is showed in Figure 8 the

terms that satisfy the searching are highlighted in the

screen. This fact allows that all these terms are

classified at the same time, making easier the

classification activity.

In Figure 8 all the terms that use “Validity” were

highlighted when the user classified that term.

Figure 8: Terms highlighted after a searching.

As it happens during the refinement activity, the

color of the boxes becomes fade as the terms are

classified. In Figure 9 all the boxes have a fade

color. This visual effect allows that the user

identifies, quickly, the terms that were defined the

ones that were not.

After all the terms were classified, the next step

provided by ONTOP-Tool – Step 3 – corresponds to

the button “Define hierarchy” that should be used to

organize, in an hierarchical way, the ontology

classes defined in the previous step. This

functionality uses a drag-and-drop interface which

facilitates this operation.

Figure 9: Visualization after the definition of all the terms.

Again, considering the ESE domain, the initial

organization of the classes is presented in Figure 10.

By means of the drag-and-drop resource the user can

reach the organization showed in Figure 11 in a

friendly way.

Figure 10: Initial hierarchy of the ontology classes.

Another resource provided by ONTOP-Tool is

available through the Step 4 that corresponds to the

button “Define relationships” of Figure 4. The screen

related to this functionality is presented in Figure 12.

In this interface, the classes defined in Step 2 are

presented on the left and on the right side of the

screen. Between them it is presented a list of

properties. These properties can be provided by the

ONTOP-Tool or can be defined by the user in this

occasion. The properties that are provided by the tool

correspond to the ones that are frequently used by

ontologies or the ones that were defined in Step 2.

ONTOP - A Process to Support Ontology Conceptualization

Figure 11: Final hierarchy of the ontology classes.

Figure 12: Screen for defining relationships.

The establishment of the relationships requires

the following actions:

(i) select a class of the right list, for example,

the Lab Package class;

(ii) select a property, for example, is_basis_for;

(iii) select a class of the left list, for example,the

Replication class;

(iv) confirm the relationship;

(v) repeat the actions (i) to (iv) until all the

relationships are established.

After these actions, the relationship “Lab

Package is_basis_for Replication” was created.

We observe that ONTOP-Tool creates

relationships of Domain-Range type. This kind of

relationship indicates that the property links the

individuals of the Domain class to the individuals of

the Range class. The other kinds of relationship that

are used in the context of ontology should be created

by the tools that support ontology development, like

Protégé-2000.

Finally, the last functionality provided by

ONTOP-Tool corresponds to the Step 5 and to the

button “Create OWL file” of Figure 4. This

functionality allows the creation of this file that

contains all the information defined till now. The

OWL file can be imported to several ontology

editors. In our research we use Protégé-2000,

versions 3.4 and 4.0.

6 CONCLUSIONS AND

FURTHER WORKS

To sum up, this paper presented the ONTOP process

which supports the conceptualization phase of an

ontology development. The tools for ontology

development identified in the field literature do not

deal with the conceptualization of the domain since

they focus on the implementation phase. ONTOP

deals with this phase and it is supported by the

ONTOP-Tool which facilitates the construction of a

collaborative glossary as well as the identification of

the ontology components.

Concerning the glossary construction, itself the

target domain should be as representative as

possible. To reach this objective it is essential that

different views and suggestions are considered. This

fact implies the involvement of different

stakeholders, especially the domain specialists that

are often geographically distributed. For the reason

we decided to use the Moodle glossary for the fact

that the Moodle environment is a free software that

provides a good set of glossary management

functionalities. By means of the Moodle

environment the glossary is easily shared and

validated by many stakeholders. Also, as it is

possible to export the glossary as an XML file, the

ONTOP-Tool provides an iteration activity that

enhances its refinement.

Another aiding support provided by the ONTOP

and ONTOP-Tool for the conceptualization phase

(which is essential for every method that supports

the ontology development, including Methontology)

is visualization. This resource was adopted in light

of two different purposes: to facilitate the glossary

refinement (for example, making easier the

identification of the definition of terms) and to

facilitate the preliminary identification of the

ontology components (for instance, using the size of

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

the boxes to select possible components of the

ontology).

An additional advantage of our proposal is the

fact that, once ONTOP-Tool can automatically

generate an OWL file at the end of the process, the

next phases of the ontology construction may be

carried out from the point where many definitions

have already been done. To continue the ontology

construction, the OWL file can be imported to a

ontology editor like Protégé-2000, among others.

All things considered, it is important to finally

point out that although we used the Experimental

Software Engineering domain to exemplify the

process and the tool, it was not our intention to

present a deeper analysis of the ontology itself, but

rather explore the ONTOP process and the ONTOP-

Tool.

In our further studies, we intend to improve the

ONTOP-Tool by adding linguistic processing so that

semantic tagging can be used to enhance the

identification phase of the ontology components.

Another functionality that we intend to add to the

ONTOP-Tool is the generation of an XMI file

(XML Metadata Interchange). This file would allow

classes, properties and relationships to be used by

UML tools.

ACKNOWLEDGEMENTS

The authors would like to thank the Brazilian

funding agencies FAPESP, CNPq, CAPES, the

institute INEP and the Project Observatório da

Educação for their support.

REFERENCES

Auvil, L.; Llorà, X.; Searsmith, D. & Searsmith, K.

(2007). VAST to Knowledge: Combining tools for

exploration and mining. In IEEE Symposium on Visual

Analytics Science and Technology, November 2007

(pp. 197-198). Philadelphia: IEEE Computer Society

Press.

Chen, C.; Kuljis, J.& Paul, R. J. (2001). Visualizing latent

domain knowledge. IEEE Transactions on Systems,

Man. and Cybernetics, Part C: Applications and

Reviews, 31(4), 518-529.

Corcho, O.; Fernández-López, M. & Gómez-Pérez, A.

(2003). Methodologies, tools and languages for building

ontologies. Where is their meeting point?. Data &

Knowledge Engineering, 46(1), 41-64.

Daconta, M. C., Obrst, L. J. & Smith, K. T. (2003) The

Semantic Web: A guide to future of XML, Web Services

and Knowledge Management. Indianapolis: Wiley

Publishing.

Fabbri, S.; Travassos, G. H.; Maldonado, J. C.; Mendonça

Neto, M. G. M. & Oliveira, M. C. F. (Eds.). (2006).

ESE Glossary: Proceedings of ESELAW’06:

Experimental Software Engineering Latin American

Workshop. Retrieved January 18, 2010, from Federal

University of Rio de Janeiro, COPPE site:

http://lens.cos.ufrj.br:8080/eselaw/proceedings/2006/pro

ceedings2006.pdf

Falbo, R. A.; Menezes, C. S. & Rocha, A. R. (1998). A

systematic approach for building ontologies. In 6th

Ibero-American Conference on Artificial Inteligence,

October 1998 (pp. 349 - 360). Lisbon, Portugal:

Springer-Verlag.

Fernández-López, M.; Gómez-Pérez, A.; Pazos-Sierra, A. &

Pazos-Sierra, J.(1999) Building a chemical ontology

using Methontology and the ontology design

environment. IEEE Intelligent Systems & their

applications, 14(1), 37-46.

Gershon, N.; Eick, S. G. & Card. S. (1998). Information

Visualization. ACM Information Visualization

Interactions, 5 (2), 9-15.

Gómez-Pérez, A.; Fernández-López, M. & Corcho, O.

(2004). Ontological Engineering. London: Springer

Verlag.

Gruber. T. R. (1993). Toward principles for the design of

ontologies used for knowledge sharing?. Knowledge

Acquisition, 5(2), 199-220.

Ichise, R.; Satoh, K. & Numao, M.(2008) Elucidating

Relationships among Research Subjects from Grant

Application Data. In 12th International Conference

Information Visualisation, July 2008 (pp. 427-432) New

York: IEEE Computer Society Press.

Johnson, B. & Shneiderman, B. (1991). Tree-maps: a space-

filling approach to the visualization of hierarchical

information structures. In 2nd Conference on

Visualization, October 1991 (pp. 284-291). San Diego,

California: IEEE Computer Society Press.

Moodle – Modular Object-Oriented Dynamic Learning

Environment. (2009). Retrieved March 10, 2009, from

http://moodle.org/

NewsMap - Application for Google News. (2009).

Retrieved December 20, 2009, from http://newsmap.jp

Noy, N. F.; Fergerson, R. W. & Musen, M. A. (2000). The

Knowledge model of Protégé-2000: combining

interoperability and flexibility. In: 12th International

Conference in Knowledge Engineering and Knowledge

Management, January 2000

(pp. 17-32). Berlin,

Germany: Springer-Verlag.

OWL - Ontology Web Language. (2009). Retrieved March

10, 2009, from http://www.w3.org/2004/OWL

Protégé-2000– Ontology Editor and Knowledge Acquisition

System. (2010). Retrieved January 06, 2010, from

http://protege.stanford.edu

TreeMap Tool. (2009). Retrieved February 7, 2009, from

http://www.cs.umd.edu/hcil/ treemap

Wohlin, C.; Runeson, P.; Höst, M.; Ohlsson, M. C.;

Regnell, B. & Wesslén, A. (2000). Experimentation in

software engineering - an introduction.Sweden:

Springer-Verlag.

ONTOP - A Process to Support Ontology Conceptualization