Ontology Reuse

Experiences from Ontology Design Pattern Selection and Integration

Birger Lantow

, Kurt Sandkuhl

1,2

and Vladimir Tarasov

University of Rostock, 18051 Rostock, Germany

School of Engineering, Jönköping University, P.O. Box 1026, 551 11 Jönköping, Sweden

Keywords: Ontology Design Pattern (ODP), Ontology Alignment, ODP Selection, ODP Integration, Ontology

Engineering, Ontology Reuse.

Abstract: While the main purpose of Ontology Design Patterns (ODPs) is to support the process of ontology

engineering, they can also be used to improve existing ontologies. This paper has a focus on ODP selection

and integration for ontology improvement. Based on the case of the ExpertFinder ontology, which allows

for competency description of researchers, selection and integration of ODP is investigated with an

explorative view. The current state of ODP selection strategies is discussed and problems arising during

integration of ODP are shown. On this base, suggestions for improvements are made. Although this study

deals with the integration into an existing ontology, most of the assumptions and suggestions are also valid

for the general case of ODP usage.

1 INTRODUCTION

Due to the increasing use of ontologies in industrial

applications at larger scale, ontology construction

and ontology evaluation have become a major area

of ontology engineering. The aim is to efficiently

produce high quality ontologies as a basis for

knowledge management, semantic web applications

or enterprise systems. Despite quite a few well-

defined ontology construction methods and a

number of reusable ontologies offered on the

Internet, efficient ontology development continues

to be a challenge, since this still requires a lot of

experience and knowledge of the underlying logical

theory.

Ontology Design Patterns (ODP) are considered

a promising contribution to this challenge. In 2005,

the term ontology design pattern in its current

interpretation was mentioned by Gangemi (2005)

and introduced by Blomqvist and Sandkuhl (2005).

Blomqvist defines the term as “a set of ontological

elements, structures or construction principles that

solve a clearly defined particular modelling

problem“ (Blomqvist 2009). Ontology design

patterns are described as encodings of best practice,

which reduce the need for extensive experience

when developing ontologies. Using ODPs, less

experienced engineers can apply the well-defined

solutions provided in the patterns when creating

ontologies.

Gangemi and Presutti (2009) discuss the

different types of ODP under investigation with their

differences and the terminology used. The two types

of ODP probably receiving most attention are logical

and content ODP. Logical ODP focus only on the

logical structure of the representation, i.e. this

pattern type is targeting aspects of language

expressivity, common problems and misconceptions.

Content ODP offer actual modelling solutions within

an application domain and are often instantiations of

logical ODP. Due to the fact that these solutions

contain actual classes, properties, and axioms,

content ODP are considered by many researchers as

tailor-made for a specific domain, even though the

domain might focus on general issues like ‘events’

or ‘situations’. This paper has its focus on the use of

content ODPs. Platforms offering ODP currently

include the ODP wiki portal initiated by the NeOn-

project (http://ontologydesignpatterns.org) and the

logical ODPs maintained by the University of

Manchester (http://www.gong.manchester.ac.uk/

odp/html/).

The contributions of this paper are (1) a

discussion of approaches for integrating ODP during

ontology engineering, (2) experiences from the

selection and integration of ODP in the ExpertFinder

Lantow, B., Sandkuhl, K. and Tarasov, V..

Ontology Reuse - Experiences from Ontology Design Pattern Selection and Integration.

In Proceedings of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2015) - Volume 2: KEOD, pages 163-170

ISBN: 978-989-758-158-8

163

case, and (3) recommendations regarding

documentation and management of ODP. The paper

extends previous work on ontology selection

strategies (Lantow et al., 2013).

The remaining part of the paper is structured as

follows: Section 2 discusses strategies for ODP

integration. Section 3 describes the process of ODP

selection and integration on the case of the

ExpertFinder ontology (see Lantow et al., (2013)).

The final section 4 summarizes the experiences and

gives recommendations for a better documentation

and management of ODP.

2 APPROACHES FOR ODP

INTEGRATION

While the selection of ODP has already been

intensively discussed in earlier work (Lantow et al.,

2013), approaches to ODP integration are in focus of

this work. This forms the base for a practical

evaluation of selection and integration approaches in

section 3 as well as a view on the influences

between the steps of ODP usage.

After selecting an ODP, it has to be integrated

into the ontology under construction in its current

state. Generally, there are two approaches to ODP

integration. First, the ODP can be considered as a

guideline how to model certain aspects of a domain.

This view on ODP is used by the ontology engineers

in Hammar (2012). The selected ODP were only

used to give an orientation for modelling. The OWL

code provided with the ODP was not applied, but the

given structures were remodelled manually. This

approach has certain drawbacks:

1. Time savings by using ODP as predefined

building blocks cannot be achieved.

2. Existing alignments of ODP cannot be used

because they are bound to the specific OWL

implementation (IRIs)

3. Considering ODP as tested, best practice

components, the manual remodelling may

introduce errors

4. Some semantic information of the patterns is not

included in their documentation but in the OWL

parts; there may be anonymous class definitions

or additional axioms. Often, these are not fully

considered, when doing manual remodelling.

Nevertheless, this approach fosters the knowledge

about the consequences of ODP integration by the

ontology engineers.

The second approach aims at the benefits of

having ODPs as building blocks for ontology

creation. The steps for using the predefined OWL

implementations for integration are first described in

(Presutti and Gangemi, 2009) and included in the

XD-Method for ontology engineering (Presutti et al.

2009). According to XD the integration consists of a

first step of importing a selected ODP into the

ontology under construction. Then four cases can be

distinguished (Presutti and Gangemi, 2009):

1. Precise Matching: The ODP is directly usable.

2. Broader Matching: The ODP intent is more

general than the local domain problem. In this

case, there should be a search for more

specialized ODP or a specialization step for the

imported ODP is necessary.

3. Narrower Matching: The ODP intent is more

specific than the local domain problem. In this

case, there should be a query for more general

ODP or a generalization step for the imported

ODP is necessary.

4. Partial Matching: The ODP intent only partially

matches the local domain problem. The ontology

specification has to be decomposed in order to

find parts that completely match ODP in the

repository. This way several ODP have to be

imported (specialized) and composed.

The task of composing ODP will commonly be a

step after import/specialization because it is unlikely

to have one ODP covering all requirements. Thus,

either several ODP need to be imported or parts of

the ontology stem from other sources. Composition

can be done by the use of relations (Object

Properties) and additional concepts in order to

connect the parts of the ontology. Furthermore,

specialization is seen as a task which commonly has

to be performed on ODP integration. Since ODP

should fit to a wide range of applications regarding

their characteristics as patterns, they tend to be more

general than local domain ontologies.

A last step seen in conjunction with ODP

integration is test and fix. Presutti et al., (2009)

suggest defining unit tests. First, the competency

questions (CQ) are transformed into formal queries

on the ontology (e.g. using SPARQL); second, some

facts are added; and finally the queries are executed

as a unit test. The unit tests are repeated with each

integration step.

The XD toolset provides automated support for

ODP integration. It contains an ODP browser, a

wizard for ODP integration by specialization and a

tool for quality assurance which helps to discover,

among other issues, usability and reasoning

problems by checking the existence of annotations

and disjointness statements.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

164

Another approach using predefined ontology

constructs as building blocks for ontology creation

stems from Iannone et al., (2009). However, this

approach does not support the direct use of OWL-

coded ODP. Instead, Ontology Pre-Processing

Language (OPPL) is used to code the ODP. OPPL is

a declarative manipulation language for ontologies.

Thus, the necessary ontology changes to integrate an

ODP can be described. The approach provides two

benefits. First, OPPL code can be used to document

the ODP usage in an ontology. Second, the

consistent use of the ODP can be assured by using

OPPL statements for instance creation. Like OWL-

coded ODP, OPPL patterns are subject to

specialization, generalization and composition but in

contrast OPPL patterns themselves have to be

changed by these operations. So far, no way has

been described how to document these changes. This

and the lack of tool support regarding the reuse of

already existing OWL-coded patterns are the reasons

for not considering OPPL further on.

3 EXPERIENCES IN ODP

SELECTION AND

INTEGRATION

This section reports on experiences in ODP selection

(section 3.1) and ODP integration (section 3.2) and

derives recommendations. The experiences are

based on the ExpertFinder task as described in

Lantow et al., (2013).

3.1 ODP Selection

Selection of ODP for improving the ExpertFinder

ontology has been subject of studies before.

Hammar et al. (2010) suggest using the “Information

Realization“ and the “N-ary Participation” pattern.

Internal course material at Rostock University

considers the construction of an ontology that fulfils

the requirements of the ExpertFinder task based on

“Information Realization”, “Participant Role”,

“Time indexed Participation”, and “Topic” pattern.

However, the process of ODP selection is not

documented in the mentioned sources. Lantow et al.

(2013) fill the gap and describe the selection process

of ODP for improving the ExpertFinder ontology.

The process followed the steps shown in Lantow et

al., (2013):

1. Filter by Domain: No directly matching domain

was discovered in the repository. The domains

“General”, “Parts and Collections”, and

“Management” were considered as partly fitting.

The domain filtering led to a reduction of ODP

candidates from 97 to 72.

2. Filter by Requirements: ODP “Intent” and

“Competency Questions” have been matched

against the general requirements of the

ExpertFinder ontology. Concentrating just on the

CQ of the ExpertFinder ontology did not seem

worthwhile because of the big difference

between the abstraction levels. The number of

candidate ODP was reduced to 35.

3. Filter by shared Conceptualizations: No direct

matches were identified. However, considering

ODP “Scenarios” and “Solution Description”

revealed at least overlapping conceptualizations

with the ExpertFinder ontology for all candidate

ODPs. Thus, the number of candidate ODP

remained 35.

4. Filter by compatibility: Manual comparison of

candidate ODP and the ExpertFinder ontology

reduced the number of candidates to 14.

Checking the compatibility of the patterns

themselves revealed that some of them were

semantically incompatible and others were

identified as integrated parts of other ODP. At

the end of the selection process, five candidate

patterns remained: “N-ary Participation”,

“Collection”, “Classification”, “Persons”, and

“Topic”.

The three independent attempts to select suitable

ODP for use in the ExpertFinder task (Hammar et

al., 2010, Rostock University course material,

Lantow et al., 2013) obviously gave quite different

results. However, the ODP catalogue and the

ontology requirements were the same. This calls for

further work on selection strategies. In general, there

are patterns for roughly the same purpose that can be

used alternatively. This is also mentioned by

(Blomqvist et al., 2009) in a general study of pattern

use. Hammar (2012) also emphasizes the importance

of pattern dependencies for pattern selection.

However, the author does not clarify what

dependencies should be considered.

Based on the ExpertFinder case and the

assumptions from the general work on ODP use, the

following recommendations can be made in order to

improve support for pattern selection:

1. Assure complete ODP description within the

given ODP metadata schema (several ODP are

not well documented (Hammar, 2012) (Lantow

et al., 2013)). Although CQ are missing only in

few cases, additional requirements may be

derived by the use of other documentation items

Ontology Reuse - Experiences from Ontology Design Pattern Selection and Integration

165

(step 2 of the selection process by Lantow et al.,

(2013)) and fitting ODP may be identified.

2. Provide a taxonomy of pattern domains. The

result would be a controlled vocabulary for the

domain description. Furthermore, sub-trees in the

taxonomy may be excluded early in the ODP

selection process. This would reduce the effort

spent in step 1 of the pattern selection process

proposed by Lantow et al., (2013).

3. Provide specialization and composition relations

between patterns. This can help to identify

incompatibilities and helps to find

specializations/generalizations. This would

reduce effort spent in step 4 of the pattern

selection process. Additionally, effort for ODP

integration (see section 3.2) may be reduced

since finding a more specialized ODP possibly

avoids ODP specialization in the ODP

integration step.

4. Name alternative Patterns and provide

information for decision between alternatives.

There are several patterns that fit the same

requirements regarding the knowledge to be

covered. The diversity of patterns that have been

proposed by the different sources for the

ExpertFinder task proves this. For a pur-poseful

selection, additional requirements have to be

considered (world view, compa-tibility, query

performance, etc.)

5. Include descriptions of common mistakes in

patterns use (Hammar, 2012)

6. Structure the pattern catalogue by architecture

tiers (Hammar, 2012)

The suggestions above are a starting point for an

improved support for (semi-automatic) pattern

selection. However, as stated in Lantow et al.,

(2013), additional aspects could also be taken into

account. There are for example potential quality

improvement and required effort for pattern

integration. This in addition with the question for

current support of ODP integration leads to the

discussion of experiences from ODP integration.

3.2 ODP Integration

Previous work on ODP integration presented

experiences only on a very coarse level. Blomqvist

et al., (2009) report generally an increase in usability

due to annotations, a better coverage of ontology

requirements regarding satisfied CQ and a decrease

of domain coverage regarding terminology. Hammar

et al., (2010) describe an incompatibility of the “N-

ary Participation” pattern with other ontologies that

have been integrated for reuse in the ExpertFinder

case. Some of the concepts were overlapping and the

pattern had to be modified for integration.

At a knowledge modelling tutorial at Rostock

university “Time-indexed participation” and

“Participant Role” ODP were chosen for integration

in the ExpertFinder case. This revealed that both

share some object properties with equivalent

semantics but different names. Four “Equivalent to”

statements had to be added:

“Event included in” ≡ “is event included in”

“Object included in” ≡ “is object included in”

“Object participating” ≡ “includes object”

“Participating in event” ≡ “includes event”

Furthermore, a specialization of both patterns in

order to combine their characteristics had to be

created. Specialization in this case required also the

consideration of anonymous super classes. While the

“Time-indexed participation” is an n-ary relation of

at least 3 concepts it should be at least 4 concepts if

a role is added.

After having these experiences from previous

studies and practical application of ODP in the

ExpertFinder task, a more detailed qualitative

assessment of the integration task has been

conducted. We separated the pattern selection and

the pattern integration part. Furthermore, the goal

was to gain some knowledge regarding the

integration of ODP in existing ontologies. The

patterns resulting from the proposed pattern

selection strategy (see section 3.1) have been

provided to an ontology engineer who had created

an ontology for the ExpertFinder task without ODP

before. This expert had to perform the integration

task based on the methodology described in section

2. He was familiar with the methodology and the

ontology he had created. Thus, we were able to

combine the findings with those of previous studies

and to avoid bias resulting from poor knowledge of

domain and methodology.

The classes of the resulting ontology including

the later ODP integration are shown in figure 1. A

semi-structured interview has been performed

afterwards in order to evaluate the support and

possible problems in the integration step of ODP

use. The interview contained the following questions

and results:

Q1: How did the ODP documentation support the

integration-process of the ODP?

So far, ODP documentation has only been

considered for ODP selection methodology.

However, documentation in software engineering is

also important for maintenance and integration.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

166

Thus, it is interesting whether this is also true for

ODP use.

The expert generally mentioned the big

differences between ODP regarding quality and

completeness of documentation. Graphical

visualizations of the ODP have been considered

helpful for ODP integration. They usually help to

grasp the intention of the pattern much faster.

However, the diagrams did not contain enough

information regarding concept identification. While

concepts of different ODP had the same label in

their visualization, they had different IRIs in their

OWL implementation. This required additional

attention in the integration process (see below).

Q2: Was it possible to use the ODP-templates?

This question aimed at the general applicability of

the proposed XD approach of ODP integration (see

section 2).

The expert was able to import all suggested ODP

OWL implementations from the repository to the

ontology.

Q3: Were changes to the ODP necessary? What

changes? Why? How many?

This question aims at exceptions from the XD

integration steps. Generally, changes to the ODP

OWL implementations contradict the approach.

ODP integration should be possible by just adding

alignment information using specialization

(generalization) and composition (see section 2).

The Persons ODP had to be changed for import.

There were contradictions caused by the pattern

based on the knowledge base. The problem was the

following statement:

persons:SocialPerson : [persons:actsThrough

owl:cardinality 1 owl:Thing]

It was either inconsistent with the functional data

properties or led to inferences as (same individuals):

expert_OhgA owl:sameAs expert_LiFe, expert_LunM

This does not reflect the real world. The following

fix had to be implemented:

persons:SocialPerson : [persons:actsThrough

owl:minCardinality 1 owl:Thing]

Another problem was the use of different IRIs in

ODP for the same concepts. The following

statements had to be added:

description:Concept owl:equivalentClass

classification:Concept

topic:Concept owl:equivalentClass classification:Concept

A third problem that occurred was the

incompatibility of the Topic and the Classification

patterns. Both patterns have been identified as

suitable to reflect research fields. Making the

ResearchField class a specialization of Topic:Topic and

Classification:Concept, lead to an inconsistent

ontology because both classes are defined as

disjoint. Three case categories can be identified

based on the described necessary changes:

1. ODP axioms are incompatible with the

constructed ontology (case 1)

2. ODP are not aligned with each other. (case 2 and

3. ODP are incompatible themselves. (case 4)

The integration of 5 ODP resulted in 4 problem

cases that had to be solved by changes and additions

to the original ODP OWL implementations. Figure 1

shows the inheritance cycles for the differently

coded Concept classes. In this case the ODP use even

results in a decrease of ontology quality.

Q4: What new relations/object properties had to be

introduced in the new Ontology besides the ones that

were already in ODP and ExpertFinder? Why? How

many?

This question aims at the specialization

(generalization) and composition steps of the XD

methodology. These steps generally lead to

additional relations besides those directly imported

as ODP. In the case of an already existing ontology

just additional inheritance may be necessary.

Furthermore, in order to use the expressivity of ODP

new object property assertions may have to be made

for existing instances.

According to the Expert, no additional object

properties had to be defined. Integration has been

done by ODP specialization:

• Collection was added as a superclass of

EducationalProgramme.

• classification:concept was used as a superclass of

ResearchField. classification:classifies was used on

the instances of ResearchField to relate them to

the instances of Degree, Expert, and Projects.

• Expert was made a subclass of

persons:NaturalPerson, Position and

UniversitySchool - subclasses of

persons:SocialPerson. persons:actsFor related

experts to positions as well as positions to the

school.

• Project and Course were made subclasses of

participation:Event. Instances of projects/courses,

experts and time intervals were related through

participation:participationIncludes.

Overall, 7 new inheritance relations and 265 object

Ontology Reuse - Experiences from Ontology Design Pattern Selection and Integration

167

property assertions have been added.

Q5: What new classes had to be introduced in the

new Ontology besides the ones that were already in

ODP and ExpertFinder? Why? How many?

This question aims at the specialization

(generalization) and composition steps of the XD

methodology. It might be necessary to create new

classes here. In the case of an already existing

ontology just additional inheritance may be enough

since existing classes may serve as specializations of

ODP classes for example. New class definitions are

only needed as long as appropriate classes are not

yet defined. Seldom, the composition step requires

new class definitions (see section 2).

In the ExpertFinder case no new classes had to

be added.

Q6: What new instances had to be introduced in the

new Ontology besides the ones that were already in

ODP and ExpertFinder? Why? How many?

Considering the possibility to discover new

requirements during ODP selection as described in

Lantow et al., (2013), it might be necessary to add

new instances to an existing knowledge base. This

question tries to assess the validity of this

assumption.

When the Classification ODP was added, 173

ResearchField instances had to be added as well. The

taxonomy of research fields had been created by a

class hierarchy but the ODP required instances in the

place of the used classes.

When the Persons ODP was added, four Expert

instances and two Position instances have been

added too. These instances did not exist in the

ontology before. By adding them the ontology

semantics became more fine-grained.

When the n-ary Participation ODP was added,

new instances of TimeInterval for projects and courses

have been created. This was necessary to represent

the time-related semantics aspects in a better way.

Two case categories for ODP use induced

instance creation can be identified:

1. ODP lead to new requirements and thus new

knowledge has to be collected in the field.

(general case)

2. ODP lead to a different representation of already

captured knowledge (classification case)

The number of additional instances highly depends

on the domain. Little dependency to the existing

ontology structure or the used ODP is assumed for

the general case. Thus, the number of new instances

is not considered in this qualitative study.

Q7: What obstacles did you encounter?

The earlier questions have been derived from the

description of the existing methods for ODP

selection and integration. This last question aims at

collecting issues that might not be covered this way.

No new aspects revealed based on the expert’s

answers.

The interview only reflects the experiences of

one ontology engineer. However, it already indicates

potential areas for improving the process of ODP

integration. Including previous studies on the area of

ODP integration the following suggestions can be

made:

1. In addition to the better documentation of

patterns as proposed in section 3.1, ODP in the

repository can be improved if equivalent classes

and properties are coded equivalent or identically

and if compatibility with certain “standard”

ontologies is assured. The answer to Q3 of the

interview showed that there were some

equivalence relations missing in the ODP. The

resulting effort for equivalence and alignment

discovery for the ODP user can be avoided.

Furthermore the goal of ontology quality

improvement is better supported. Inheritance

cycles or unnecessary complexity, respectively,

can be avoided.

2. Documentation of ODP can be further improved

if equivalent parts of several ODP are made

explicit. This way the ODP user is aware of

equivalent concepts and ODP parts. A better

understanding of the ODP structure is assumed.

This revealed to be problematic looking into the

answers to Q1 and Q3.

3. The integration step includes in addition to

specialization (generalization) and composition

also ontology alignment and adaptation of

ontology axioms. For example, cardinalities or

anonymous classes need to be considered. It has

been shown that the integration steps as

described in the XD method are not enough.

Axioms and anonymous classes have to be

considered as well regarding reasoning of new

facts and consistency. This problem showed in

the knowledge modelling tutorial and in the

answer to Q3 as well.

4. The test step should not only consider querying

but also consistency. The case study has shown

that ODP use may also lead to inconsistencies

(see answer to Q3). These can hardly be

discovered just by SPARQL queries.

5. Tools should allow for integration of ODP parts

(see also Hammar (2014)) since sometimes parts

of the ODP are already covered by existing

classes.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

168

Figure 1: ExpertFinder Ontology after ODP integration - Inheritance Structure.

6. More integration related aspects can be included

into pattern selection. For example: necessity of

additional instances in order to make ODP more

usable or number of axioms that have to be

evaluated during ODP integration.

4 CONCLUSION

Although this work is rather explorative, it covers

recent approaches to ODP reuse and shows

important challenges for a broad use of ODP in

ontology engineering based on practical experiences.

Furthermore, potential areas for improvement are

discussed. With a rising number of ODP available in

the repository, automatic or semi-automatic pattern

selection becomes more important. So far existing

approaches provide only little help.

Improvements regarding recall have been

achieved (e.g. by Hammar (2014)) but precision is

poor. However, if more and more ODP have a

similar domain and similar requirements, the number

of alternatively usable ODP increases. Thus,

precision might not be that important in the future.

Structured ODP selection processes might help to

guide the ontology engineer. However, improved

tool support and improved ODP documentation is

needed. Improved ODP documentation includes

using the current metadata schema more extensively

and also the addition of new metadata, like pattern

compatibility (cf. section 3.1 and 3.2). Also an

improvement of the patterns themselves regarding

compatibility with existing ontologies, the treatment

of equivalent classes and properties as well as the

fitness of ODP axioms for reuse needs to be

addressed.

The structure of the interview can be used for

further investigations on ODP use and ontology

reuse in general.

REFERENCES

Blomqvist E., Sandkuhl K. (2005) Patterns in Ontology

Engineering – Classification of Ontology Patterns.

Proc. 7th International Conference on Enterprise

Information Systems, Miami, USA, May 2005.

Blomqvist, Eva (2009) Semi-automatic Ontology

Construction based on Patterns. PhD thesis, Linköping

University, Department of Computer and Information

Science, 2009.

Blomqvist, Eva; Gangemi, Aldo; Presutti, Valentina

(2009): Experiments on pattern-based ontology

design. In: K-CAP, pp. 41–48.

Brank, J., Grobelnik, M., Mladenić, D., 2005. A survey of

ontology evaluation techniques, in: Proceedings of the

Conference on Data Mining and Data Warehouses

(SIGKDD, 2005). Ljubljana, Slovenia.

Corcho, O., Roussey, C., Vilches-Blázquez, L. M., &

Perez Dominguez, I. (2009). Pattern-based OWL

ontology debugging guidelines. CEUR Workshop

Proceedings.

Duque-Ramos, A., Fernandez-Breis, J.T., Stevens, R.,

Aussenac-Gilles, N., 2011. OQuaRE: A SQuaRE-

Ontology Reuse - Experiences from Ontology Design Pattern Selection and Integration

169

based approach for evaluating the quality of

ontologies. Journal of Research and Practice in

Information Technology 43, 159.

Gangemi, A. (2005) Ontology design patterns for semantic

web content. In The Semantic Web ISWC 2005, Vol.

3729, Lecture Notes in Computer Science. Springer,

2005.

Gangemi, Aldo; Catenacci, Carola; Ciaramita,

Massimiliano; Lehmann, Jos (2005): A theoretical

framework for ontology evaluation and validation. In:

SWAP, vol. 166..

Gangemi, A. and V. Presutti (2009) Ontology design

patterns. In Handbook on Ontologies, 2nd Ed.,

International Handbooks on Information Systems.

Springer, 2009.

Hammar, K., Tarasov, V. & Lin, F. (2010). Information

Reuse and Interoperability with Ontology Patterns and

Linked Data: First Experiences from the ExpertFinder

Project. In: Witold Abramowicz, Robert Tolksdorf,

Krzysztof Wecel (Ed.), Business information system

workshops: BIS 2010 International Workshops, 3rd

Workshop on Information Logistics and Knowledge

Supply in conjunction with 13th International

Conference on Business Information Systems (pp. 168-

179). Berlin: Springer, LNBIP 57

Hammar, Karl (2012): Ontology Design Patterns in Use -

Lessons Learnt from an Ontology Engineering Case.

In Eva Blomqvist, Aldo Gangemi, Karl Hammar,

María del Carmen Suárez-Figueroa (Eds.):

Proceedings of the 3rd Workshop on Ontology

Patterns, Boston, USA, November 12, 2012: CEUR-

WS.org (CEUR Workshop Proceedings, 929).

Hammar, Karl (2014): Ontology Design Patterns:

Adoption Challenges and Solutions Web Enterprise

Adoption and Best Practice and Second International

Workshop on Finance and Economics on the Semantic

Web Co-located with 11th European Semantic Web

Conference, WaSABi-FEOSW@ESWC 2014,

Anissaras, Greece, May 26, 2014.

Hong, Tzung-Pei; Chang, Wen-Chang; Lin, Jiann-Horng

(2005): A Two-Phased Ontology Selection Approach

for Semantic Web. In Rajiv Khosla, Robert J. Howlett,

Lakhmi C. Jain (Eds.): Knowledge-Based Intelligent

Information and Engineering Systems, vol. 3684:

Springer Berlin Heidelberg (Lecture Notes in

Computer Science), pp. 403–409. Available online at

http://dx.doi.org/10.1007/11554028_56.

Iannone, L., Rector, A. & Stevens, R. (2009). Embedding

knowledge patterns into owl. In The Semantic Web:

Research and Applications (pp. 218-232). Springer

Berlin Heidelberg.

Lantow, Birger; Sandkuhl, Kurt; Tarasov, Vladimir

(2013): Selecting Content Ontology Design Patterns

for Ontology Quality Improvement Knowledge Supply

and Ontologies in Information Systems, Warsaw,

Poland, September 23rd, 2013.

Maedche, A., Staab, S., 2002. Measuring Similarity

between Ontologies, in: Gómez-Pérez, A., Benjamins,

V.R. Eds., Knowledge Engineering and Knowledge

Management: Ontologies and the Semantic Web.

Springer Berlin Heidelberg, Berlin, Heidelberg, pp.

251–263.

Mikroyannidi, E., Manaf, N. A. A., Iannone, L. & Stevens,

R. (2012). Analysing Syntactic Regularities in

Ontologies. In OWLED (Vol. 849).

Poveda-Villalón, M., Suárez-Figueroa, M. C. & Gómez-

Pérez, A. (2012). Validating ontologies with OOPS!.

In Knowledge Engineering and Knowledge

Management (pp. 267-281). Springer Berlin

Heidelberg.

Presutti, Valentina; Gangemi, Aldo (2008): Content

Ontology Design Patterns as Practical Building Blocks

for Web Ontologies. In Qing Li, Stefano Spaccapietra,

Eric S. K. Yu, Antoni Olivé (Eds.): Conceptual

Modeling - ER 2008, 27th International Conference on

Conceptual Modeling, Barcelona, Spain, October 20-

24, 2008. Proceedings: Springer (Lecture Notes in

Computer Science, 5231), pp. 128–141.

Presutti, Valentina; Daga, Enrico; Aldo Gangemi;

Blomqvist, Eva (2009): eXtreme Design with Content

Ontology Design Patterns with the 8th International

Semantic Web Conference (ISWC-2009), Washington

D.C., USA, 25 October, 2009.

Sabou, Marta; Lopez, Vanessa; Motta, Enrico (2006):

Ontology Selection for the Real Semantic Web: How

to Cover the Queen’s Birthday Dinner? In Steffen

Staab, Vojtech Svátek (Eds.): Managing Knowledge in

a World of Networks, 15th International Conference,

EKAW 2006, Podebrady, Czech Republic, October 2-

6, 2006, Proceedings: Springer (Lecture Notes in

Computer Science, 4248), pp. 96–111.

Vrandečić, D., Sure, Y., 2007. How to Design Better

Ontology Metrics, in: Franconi, E., Kifer, M., May,

W. Eds., The Semantic Web: Research and

Applications. Springer Berlin Heidelberg, Berlin,

Heidelberg, pp. 311–325.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

170