SELECTING AND STRUCTURING SEMANTIC RESOURCES TO

SUPPORT SMES KNOWLEDGE COMMUNITIES

António Lucas Soares

INESC Porto, Rua Roberto Frias, Campus da FEUP, 4200 Porto Portugal

University of Porto, Faculty of Engineering, Rua Roberto Frias s/n, 4200 Porto Portugal

Manuel Moreira da Silva

INESC Porto, Rua Roberto Frias, Campus da FEUP, 4200 Porto Portugal

ISCAP - IPP, Rua Jaime Lopes Amorim, s/n, 4465 S. Mamede de Infesta Portugal

Dora Simões

INESC Porto, Rua Roberto Frias, Campus da FEUP, 4200 Porto Portugal

University of Aveiro, ISCAA, R. Associação Humanitária dos Bombeiros de Aveiro, 3811 Aveiro Portugal

Keywords: Semantic resources, Knowledge communities, Ontology engineering, Ontology reuse/integration, KMS.

Abstract: Knowledge management intrinsically involves communication and information sharing, which can be

strongly affected by the context in which it is viewed and interpreted. This situation gets worst when

complex domains are considered, as it is the case of the Construction Industry domains. The development of

ontologies to unify and to put into context the different concepts and terms of the sometimes rather

traditional and locally coloured construction industry domains is a necessary step to avoid

misinterpretations and inefficient communication. The KNOW-CONSTRUCT project decided, as an

approach to this task, to re-use, as far as possible, existing ontologies, classification systems and other

semantic resources in order to develop a system that may come to contribute to standards and to the

integration, management and reuse of the area specific knowledge via a common knowledge base in order to

consolidate and provide access to integrated knowledge, making community emergent knowledge a

significant added value. It aims at developing a methodology of common Construction Industry Knowledge

(CIK) representation applicable to large sets of SMEs in the construction industry as a basis for the

establishment of a knowledge community.

1 INTRODUCTION

Building and construction companies have to

continuously renew their working habits in order to

face an increasing competitive environment where

flexibility and adaptability to change are the obliged

route to success.

The main challenge is to provide a cost-effective

solution for the two main problems:

1. Construction industry (particularly SMEs)

urgently needs radical improvements of

communication with customers in order to

provide better product support and services. The

innovative forms of communications and

relationships among SMEs and their customers

are increasingly important in order to improve

the market share and/or survival chances in the

"new economy era".

2. To respond to ever increasing customer

requirements it is increasingly necessary to

establish a closer co-operation (particularly

among SMEs) within this sector, aiming at

assembling alliances of SMEs into integrated

teams that will genuinely align with challenging

performance targets.

270

Lucas Soares A., Moreira da Silva M. and Simões D. (2006).

SELECTING AND STRUCTURING SEMANTIC RESOURCES TO SUPPORT SMES KNOWLEDGE COMMUNITIES.

In Proceedings of the Eighth International Conference on Enterprise Information Systems - AIDSS, pages 270-277

DOI: 10.5220/0002465602700277

 SciTePress

Taking into account these needs the the Know-

Construct (KC) EU project

aims to develop an

Internet Platform for Knowledge-based Customer

Needs Management and for Support to Knowledge

Communities of SMEs in Construction Industry. It

involves professional associations of the sector as

they may provide an ideal environment to

disseminate innovative solutions.

The goal of KNOW-CONSTRUCT (KC)

(www.know-construct.com

) is to provide a common

internet-based platform for SMEs from the

construction sector to achieve an effective

combination of two general functionalities:

1. Customer Needs Management (CNM) System: a

decision making support system regarding the

products characteristics, applications and other

consultancy services for SMEs customers

applying the "web enabled dialogue", and

2. Knowledge Communities Support (KCS) System:

a system for SMEs to support a form of co-

operation through the creation of Knowledge

Communities of SMEs in Construction Industry.

Figure 1: Know-Construct basic concept.

The system supports the integration,

management and reuse of the area specific

knowledge via a common knowledge base. The

platform will be owned by associations, offering to

members (SME) an area to establish specific

customer relationship management (using the KC

individual CNM systems) that take benefit of a

knowledge community supported by the KCS

System.

For a successful solution it was necessary to

establish a multidisciplinary approach combining

research activities in different areas, such as

knowledge representation, particularly creation of

the adequate ontologies and their structure, effective

forms of on-line interaction among customers and

SME (eConsultancy - web enabled dialogue with

EU - COLL-CT-2004-500276

expert/technical advisor), as well as product

classification systems for this sector, new forms for

a representation of the CI knowledge (e.g.

experience-based), investigation of efficient

approaches for training, etc.

This paper describes a fundamental phase of the

project which was the selection and structuring of

the semantic resources (ontologies, classifications,

thesauri), that form the cornerstone of the KC

system. The rest of the paper is as following: section

2 describes briefly in the concept and functional

architecture of the KC system; section 3 describes

the process of selecting the relevant knowledge

sources for the KC ontology construction and

integration; section 4 describes the KC ontology

structure and gives examples of its content.

2 SME KNOWLEDGE

COMMUNITY SUPPORT

SYSTEM

The relationship between organizations and

communities, including their implication in the

organization structure, IT systems and business

model, depends on the nature of the value derived of

them. This area has not been sufficiently mapped

before, but [Cornejo 2003] explores it very well

starting with the nature of knowledge, the translation

into value for individuals and organizations, the

definition of community taxonomies and its

influence on organizational drivers. KC addresses

this specific point: how to derive the maximum

value to individual SMEs from a knowledge

community where those companies participate.

2.1 Models for Knowledge

Communities

KC general objective is to improve the relationship

of CI SME's with their customers by providing the

later an innovative support regarding information

and knowledge about products, processes and

associated issues. This is achieved through

specifically developed ICT tools, in particular tools

that support the formation and operation of SME's

knowledge communities, fostering an improved

collaboration aimed at generating broader and more

accurate knowledge to be used in satisfying

customers.

The broader context of this development are the

sector’s associations that provide, besides others,

some sense of grouping to the participant SME's.

This knowledge community (Construction Industry

SELECTING AND STRUCTURING SEMANTIC RESOURCES TO SUPPORT SMEs KNOWLEDGE COMMUNITIES

271

Knowledge (CIK) community) can be considered as

a hybrid of community of practice and interest,

which is Organization-sponsored (Porter, 2004).

From one side, company employees as individuals

should see a direct utility to their particular jobs

when participating in the CIK community. This

direct utility comes into light when an employee

(and consequently the company) realizes that, when

solving a problem to an important customer (using

the CNM system), the information/knowledge used

to reach the solution comes also from the

contribution of the other community members.

Nevertheless, not all the activities can be tracked

to a causal benefit to the SME. For example, some

chatting between two employees exchanging a

specific professional experiences or a report in a

news or blog item by another employee of a concern

regarding the performance of a construction

material, are activities that make sense in a

community but cannot be assigned of a concrete

immediate value for the organization.

Based on work of de Vries et al. 2004, we

synthesize the following characteristics of the CIK

Community:

• The goal is to develop and exploit knowledge

about construction area.

• There are continuous interactions between

participants in order to meet these goals.

• Information and communication processes are

continuously made explicit.

• It adds value to the participants (professionals

and customers).

• The online meeting place is usable.

• The culture focuses on the needs of the

participants as the route to high performance;

involvement and participation create a sense of

responsibility and ownership and, hence, greater

commitment to the community.

• The context is highly complex and very unstable,

and CIK Community will have to continuously

comply with the expectations of its participants

and their context of use.

The main question here is how to implement and

make use of online knowledge communities in order

to meet these goals and expectations. As an answer

to this problem, we propose to adopt and adapt the

reference model developed by [de Vries et al. 2004]

that describes factors that affect the implementation

and use of an online knowledge community as a type

of networked organizational communication.

Following this explanation, CIK Community

will consist of three elements: professionals, on-line

professional meeting place, and organizational

context. We see the implementation of CIK

Community as a continuing communication process,

a constant search for a fit between these three

elements:

• Professionals are all the professionals of the

construction area and belonging to the

associations of the project partnership.

• On-line Professional Meeting Place is the

online place where the members of the

community will take theirs discussions, debates,

conversations, etc.

• Organizational context is related with the

organizational goals, culture, technology, etc.

2.2 KCS System General

Architecture

As mentioned before, the KC project has a very

specific goal: to enable individual SME's to better

solve the problems of their customers. Therefore,

KCS is focused on pursuing this goal in the first

place. Although a knowledge community

encompasses, as stated before, mechanisms that

surpass this simple instrumental goal, the initial

vision of KCS is developed from this

instrumentality. This means that KCS supports CIK

community building in a broad sense, but focused

fundamentally in generating broader and deeper

knowledge to be used in managing the SME's

customers' relationship, particularly in problem

solving.

The system (KCS) provides the following three

general functions:

1. Community building tools: this part of KCS

supports the processes of community building by

providing the instruments to foster professional

interaction and socialization. Forums and

weblogs are two of such instruments and are

tailored in KCS to be strongly integrated with the

semantic structure supporting knowledge

management in KC.

2. Semantic resources management: this is a set

of infrastructural functionalities that support

information and knowledge acquisition,

organization and storage in KCS (and also

CNM). More specifically, it enable the (i)

management of classifications, thesauri and

vocabulary, (ii) the acquisition of knowledge

from digital content (including forums and

weblogs entries, web pages, etc.) both internal to

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

272

Figure 2: The Know-Construct functional architecture.

the CIK and from external sources, (iii) the

maintenance of an ontology which is the base of

knowledge representation, access and storage.

3. Knowledge resources access: creating,

searching and updating knowledge resources will

be a fundamental set of functionalities in KCS.

Although much of the community

information/knowledge will be created in

communication/interaction processes (foruns,

weblogs), there will be also the need to

create/access knowledge in a more structured

way. Digital content management and document

management are the natural approaches

regarding this issue.

This generic architecture is represented in more

detail in Figure 2.

3 SELECTION OF KNOWLEDGE

SOURCES FOR KCS IN THE

CONSTRUCTION INDUSTRY

SECTOR

The identification and selection of existing

knowledge sources was the first step to build the

semantic resources structure of KCS system.

Multiple different sources' categories like

terminologies, ontologies, international

classifications, standards, norms and regulations,

national classifications were analysed. The difficulty

of this task is well known, since the different sources

are usually designed using different theoretical

grounds and design principles.

3.1 Types and Characteristics of

Knowledge Sources

According to ISO 12006-2 (ISO, 2001) the most

widely used classifications are work sections

(mainly for specifications) and elements (mainly for

cost analysis). They are also the most widely varied,

not only in their itemization and structure but also in

the range of other purposes to which they are put. As

a result of our research, several other classifications

were identified, potentially just as important, which

have not yet been used to the same degree, e.g.

construction products and properties/characteristics.

KC project is, along with the necessary

development of classifications and taxonomies that

answer the project’s needs, re-using/integrating as

far as possible existing ontologies, classification

systems and terminologies in order to develop a

system that may, in the future, contribute to

standards. The initial interaction of the KC project

with standards issues will be to assure full

compliance of the developed solution components

with the current legal and de-facto standards in the

targeted building sector and in relevant ICT

domains.

The ontologies were developed in the areas of

product characteristics, product applications and

related consultancy services. These ontologies are

crucial for the decision making support system but

also to create uniform models for customer's access.

Standard or integrating ontologies do not exist in

SELECTING AND STRUCTURING SEMANTIC RESOURCES TO SUPPORT SMEs KNOWLEDGE COMMUNITIES

273

these areas. Furthermore, another essential

innovation, potentially contributing to Construction

Industry standards, is the development of integrating

ontologies both in the areas referred above and in

inter-enterprise interoperability. It is exploiting

proposals used or in use in other European and

international projects, in an attempt of

harmonization with the current well established

standards, also as a way to oppose the unpredictable

perennity of the stored data.

Besides CI online sites and other sources, some

of the most relevant ontologies and classifications

identified so far are: e-Cognos ontology -

Methodology, tools and architectures for electronic

consistent knowledge management across projects

and between enterprises in the construction domain;

e-Construct ontology - Electronic Communication in

the Building and Construction Industry; EPIC –

European Product Information Co-Operation;

UNICLASS – Unified Classification for the

Construction Industry; IFC Model – Industry

Foundation Classes; ICIS LexiCon – International

Construction Information Society.

The identification of these sources led to the

conclusion that part of the existing information has

some common principles and structures, mostly

because they result from European or governmental

projects which also aim to contribute to

harmonization and standardization. But, its diversity,

nevertheless, puts us before the problem of how to

adapt the selected resources according to KC

project purposes and scope and the industry

consortium predefined needs, taking also into

consideration the specific cultural and professional

context of the ontology’s development and use; the

target-audience(s); the previously defined scenarios.

Table 1: Criteria for the evaluation of knowledge sources to be used in KC.

Type of source Common Criteria Description

Origin Developer(s); type of entity (CEN, ISO, DIN, other)

Relevance for the pre-defined areas of analysis; for specific cultural and professional

tt()

Adequacy from the domain expert point of view; from the ontologist point of view

Completeness explicit in-depth coverage

Comprehensiveness domains addressed in the area

Ease of data Possibility of access and reuse (merge/integrate)

Language language(s) in which it is available, multilingual features, language

All

Current status finished, work in progress, in revision

Specific Criteria

Conce

tual ontolo

assum

tions and ontolo

ical commitment and their relation to KC

Type of concept identification of generic concepts and relationships, identification of domain

concepts and relationships

Design principles internal structure

Knowledge acquisition Quality of knowledge sources, adequacy of knowledge acquisition practices

Supported applications applications supporting the ontology codification language

Documentation type of documentation available and accessibility

Consistency consistency of the application of the relations

Ontology

Modularity which concepts are represented in which modules

Terminology purpose operational terms – functions the terminology is intended to serve

Standardized/non- implemented as standard, other type

Granularity level of complexity of the available data

Quality of the do they follow unified patterns, are simple, clear, concise, etc.

Interconnectivity to what extent is the terminology mappable to coding systems or

Precision and recall retrieval effectiveness

Normalization of content and semantics

Terminology

Responsiveness Frequency of update

Classification classification purposes and their relation to KC objectives

Conceptual framework classification assumptions and their relation to KC objectives

Classification scope Domain(s)

e of conce

ts de

ree of abstraction/s

ecificit

Classification

Previous use use in ontology projects and outcome analysis

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

274

3.2 Evaluation and Selection of

Knowledge Sources

In order to choose the knowledge sources, a

complex set of multi-criteria referring to different

aspects were established. An analysis framework

was elaborated in order to evaluate the analysed

knowledge sources in what concerns to their

suitability to the KC system and methodology.

The choice of the evaluation methods to use

depends on different factors such as:

elements and indicators previously

identified,

knowledge to be represented,

stage of system development,

 time and resources available,

type of output required,

precision / reliability desired.

Along with these criteria, subject field

specialists were consulted to analyse both the

methodologies and the resulting ontology. After

this process of identification, classification and

evaluation, and taking in consideration the general

view expressed in the CWA 15142 – European

eConstruction Ontology (EeO), from the available

sources, those which, at the moment, present the

best solutions for the purpose of reuse and

integration of information and for the development

of Know-Construct high-level ontology: LexiCon,

bcBuildingDefinitions, eCognos, IFC model, ISO

12006 - 2 and 3 - Building construction —

Organization of information about construction

works

Table 1 describes the criteria established and

their scope inspired in the works of Pinto and

Martins (2001) and Lelkin (2004).

4 ONTOLOGIES STRUCTURE

AND CONTENT

After the activities described above, we delineated

the adequate strategy in order to support the

structuring, maintenance and evolution of CIK

ontology and local ontologies. This level of

ontology management is necessary not only for the

initial development and maintenance of ontologies,

but it is essential during deployment, when

scalability, availability, reliability and performance

are absolutely critical.

As it had been defined in the KC proposal,

the methodology to be followed aims mainly at

identifying, evaluating and reusing existing

semantic resources, like ontologies in the IC area,

due to the advantages their reuse offers: simple

design, more reliable knowledge sharing and

clearer semantic representations. New

developments made in the construction industry

context or in any other will be submitted to the

above listed representative type of ontologies.

4.1 The Local/Global Semantic

Resources Approach

During the analysis described above, a perception

gained strength – to deal with the very concrete

reality of the Associations and SMEs of each

country, KC system would need to have a local

ontology that would answer the KC partners’

particular professional and cultural needs.

Figure 3: General structure of the KNOW-CONSTRUCT CIK Ontology.

SELECTING AND STRUCTURING SEMANTIC RESOURCES TO SUPPORT SMEs KNOWLEDGE COMMUNITIES

275

The development of this more specific ontology

will be based on a larger, upper level ontology – the

CIK ontology, where all the central concepts of the

Construction Industry area are structured.

The perception of the need to provide tailored

support to the users was reinforced by the

conclusions of the CWA 15142 on European

eConstruction Ontology (EeO), where it is clearly

stated that the e-COGNOS vision over the

development of a big ontology was confronted with

an unexpected reality. The end users actually

showed their preferences to use their very specific,

concise and precise taxonomies. They did not want

to handle big ontologies; rather they are perfectly

happy if their small resources are in place providing

the results they are expecting. This fact has changed

the concept of the e-COGNOS ontology: the big

ontology is in place, but it is totally customisable in

the sense that a small taxonomy with 100 concepts

can replace the big one.

KC has decided to take this fact into account and

look at this possibility as part of the standardised

way to develop ontologies in the sector, but in such

an away as to keep a common central ontological

content (structure, attributes, relations, etc.) from

where to derive the more specific ontologies.

Therefore the solution proposed is to develop an

inter-organizational KM system for Construction

Industry Knowledge Communities which will be

built upo8n distributed ontologies locally managed

and centrally integrated.

The central ontology reflect standards and

related classification schemes in the industry and the

local ontologies will account for the individualised

SME conceptual schemes, i.e. they will be strongly

related to the consortium partners' needs. This

methodology results in a need to develop the two

types of ontologies in two different moments which

leads to two main concerns: 1) how to establish

through the re-use and integration of existing

ontologies (as far as possible) an adequate domain

related ontology, as well as classification system for

this sector applicable in SMEs environment, and 2)

how to assure the continuous update/maintenance of

both types of ontologies in order to enable a long life

to the knowledge systems.

4.2 Building and Integrating the

Global Ontology

Ontologies represent shared knowledge between the

parties, and result from a shared approach to a

knowledge domain. In the case of KC, the decision

to re-use and integrate available ontologies and

classifications standards in the CI area, led to an

even greater need to develop a Know-Construct

high-level ontology that would allow the integration

of these resources. In order to better answer the

needs and purposes of the CNM and KCS systems,

this ontology (named as CIK Ontology) is integrated

in the Enterprise Ontology, as defined by the

Enterprise Project by the Artificial Intelligence

Applications Institute at the University of

Edinburgh.

The following description respects only to the

CIK Ontology that defines the domain of the CI

which can be summarized in the following sentence:

The Construction Industry involves a set of

resources (Construction Resource) that follow

certain conditions (Technical Topic) which are used

or required in a process (Construction Process) that

leads to results (Construction Result).

As such, the proposed taxonomy includes four

major domains to classify these major concepts:

• Construction Resource

• Construction Process

• Construction Result

• Technical Topic

As it can be seen the first three domains coincide

with the major themes in the ISO 12006-2 standard.

The other domain (Technical Topic) is the result of

the integration of an e-COGNOS module, further

developed so as to include issues related to the CI

that are not covered by the e-COGNOS ontology

and IFC model.

4.3 Using and Integrating Local

Semantic Resources

Other well developed resources that may be useful

for the CIK ontology definition and further

development are the country specific semantic

resources. These resources are highly developed and

have been used in previous projects in the CI area

and, in the majority of the cases; they are the result

of the work of several combined institutions and

actors interested in developing the existing standards

and classifications.

These sources are mandatory for future

developments of the local ontologies, for the specific

markets and specific Costumer Needs Management

and Knowledge Community Systems, the sources

have to be defined according to the local market

context, SMEs and the Industrial

Association/Grouping needs.

They need to be an integrant part of the local

market information resources, the SMEs internal

documentation and catalogues, internal documents

and databases (contacts, material properties,

specifications, standards, prices), and the diverse

ICEIS 2006 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

276

publications in the area of the CI, especially those

edited by the Associations or available in the

internet (portals, individual sites, online libraries,

newsgroups, etc.).

On a second stage of the ontology definition and

management – the establishment of the local more

specific ontologies - other more detailed local

specifications/structures will have to be considered

and developed taking into account the local/national

sources of knowledge and of knowledge

dissemination, also as a way to better address the

management of tacit knowledge and, in particular,

the social aspects of knowledge exchange.

As these sources are based on different

conceptual models, in order to integrate them we

must solve the problem of semantic heterogeneity

between them. From the review on different

approaches to support this task, we've decided to

follow the ontology mapping approach that allows

relating a portion of the source ontology to the target

ontology’s entities, transforming instances from the

source ontology into instances in the target ontology.

Ontology mapping approach allows transforming

information but doesn’t require the building of an

integrated view. So, although it’s more powerful

than simple inclusion, it avoids the complexity and

overhead of integrating multiple sources.

5 CONCLUSIONS AND FUTURE

WORK

Given the selection, evaluation and structuring

process described in this paper, there is an high level

of probability that the CIK ontology reflects the

standards and related classification schemes in the

industry, on the one hand, and, on the other, the

local more specific ontologies will account for the

individualized SME conceptual schemes, i.e. they

will be strongly related to consortium partners'

needs, as identified in the analysis of business case

scenarios and in the users requirement definitions.

The implemented method aims at developing a

methodology of common Construction Industry

Knowledge representation applicable for large sets

of SMEs in the construction industry as a basis for

the establishment of a knowledge community.

Because of the available time frame, the

described method was not as fine grained as we

would like it to be. Therefore, further work in this

area will be directed to detail the evaluation criteria.

Next steps will involve the definition of the

ontologies and classifications maintenance

strategies. In parallel, the first prototype validation

by the users will be setup.

REFERENCES

CEN - European Committee for Standardization; CWA

14946:2004 E, European eConstruction Framework

(EeF) - Context and Scope for “eConstruction”,

Brussels, 2004.

CEN 2003, CEN/ISSS Roadmap for addressing key

eBusiness standards issues 2003-2005, source:

CEN/ISSS eBusiness Standards Focus Group.

http://www.eeuropestandards.org/Docs/

Cornejo, M. 2003. Utility, value and Knowledge

Communities. Knowledge Communities (to be

published).

CWA3 – CEN Workshop Agreement, European

eConstruction Ontology (EeO), 2004. Documents

produced by the CEN/ISSS eConstruction Workshop,

Brussels, 2004.

De Vries, S. and P. Kommers 2004. "Online knowledge

communities: future trends and research issues."

International Journal of Web Based Communities

e-COGNOS Public Deliverables, available at

http://www.e-cognos.org/

eConstruct Public Deliverables, available at

http://www.bcxml.org/

FUNSIEC, Feasibility study for an UNified Semantic

Infrastructure in the European Construction sector –

Public documents available at www.funsiec.org

Gronau, N. and E.-M. Kern 2004. Collaborative

Engineering Communities in Shipbuilding. Virtual

Enterprises and Collaborative Networks (PROVE'O4),

Toulouse, France, Kluwer Academic Publishers.

ISO 12006-2 Building construction — Organization of

information about construction works - Framework for

classification of information, DIS version 2001.

Lima, C., A Framework to Support Interoperability among

Semantic Resources, 2004.

Maedche, A., B. Motik, et al. 2003. "Ontologies for

enterprise knowledge management." IEEE Intelligent

Systems (Intelligent Information Processing): 2-9.

Pinto, Helena S.; and Martins, João P. 2001; Ontology

Integration: How to perform the Process. Proceedings

of the international conference on Knowledge capture.

Victoria, British Columbia, C

anada.

SELECTING AND STRUCTURING SEMANTIC RESOURCES TO SUPPORT SMEs KNOWLEDGE COMMUNITIES

277