26 ◾ Rajendra Akerkar
Attraction
Human artifacts
Museum
Visitor center
Art gallery
Architecture
Stave church
Stone church
Memorial
Monument
Viewpoint
Picnic area
Natural sites
Fjord
Waterfall
Mountain pass
National park
Activity
Skiing
Ski lift
Ski tracks
Rafting
Fishing
Walking
Climbing
Cycling
Kayaking
Excursion
Guided tour
Fjord tour
Canyon tour
Glacier tour
……….
Figure2.1 Hierarchical representations of tourist attractions.
Ontology: Fundamentals and Languages ◾ 27
A thesaurus can be modeled by concept- and term-oriented models. e
International Organization for Standardization (ISO) provides two standards that
deal with thesauri: ISO 2788 for monolingual thesauri and ISO 5964 for multilin-
gual thesauri. Figure2.2 is a partial thesaurus example (sorted alphabetically) for
a typical tourism industry system. Both taxonomies and thesauri provide vocabu-
laries of terms and simple relationships. erefore, taxonomies and thesauri are
above XML, namespaces, and controlled vocabulary in the Semantic Web stack.
Accommodation Bed and breakfast
Cabin
Farm house
Hotel
Pension
Activities Climbing
Cycling
Fishing
Hiking
Kayaking
Rafting
Ski tracking
Skiing
Walking
……… ……….
Events Cultural
Sport
Excursions Canyon tour
Fjord tour
Glacier tour
Guided tour
Railway
Figure2.2 Example of thesaurus.
28 ◾ Rajendra Akerkar
However, the relationships they express are not as rich as those provided by RDF
or Topic Maps, and consequently by ontology. In general, ontology consists of a
taxonomy combined with relationships, constraints, and rules; the rules may be
used with RDF or Topic Maps.
Ontology enables us to agree upon the meanings of terms used in a precise
domain, knowing that several terms may represent the same concept (synonyms)
and several concepts may be described by the same term (ambiguity). Ontology con-
sists of a hierarchical description of important concepts of a domain and a descrip-
tion of each concept’s properties. Ontology is at the heart of information retrieval
from nomadic objects from the Internet and from heterogeneous data sources. An
address can be modeled as shown in Figure2.3.
In semantic-based information retrieval, ontology directly species the mean-
ings of concepts to be searched. XML-based systems have very limited utility in this
context unless the independent site content authors agree on the semantics of the
terms they embed in source metadata. Ontology reduces such semantic ambiguities
by oering a single interpretation resource. Furthermore, ontology can also enable
software to map and transform information stored using variant terminologies.
Some researchers adopt modeling terminology and consider ontology as a meta-
model, dened as an explicit description of the constructs and rules needed to
Deta iledAddress
Office
Apartment
Address
BuildingCategory
ing
PublicPlace
Flat
Room
Store
Road
Avenue
Street
Drive
is-a
is-a
is-a
is-a
is-a
is-a
is-a
is-a
is-a
is-a
is-a
is-a
is-a
Figure 2.3 Example of address ontology.
Ontology: Fundamentals and Languages ◾ 29
build specic models within a domain of interest. A specic model can be created
by instantiating the types and relating instances to each other according to the
relationships in the meta-model; a model of the domain; and an example of a more
general model, i.e., a meta-meta model. In order for a meta-model to act as an
ontology, three properties must hold: (1) it must be expressed in a formal language
to enable consistency checks and automated reasoning (formalization), (2) it must
be agreed upon by a community (consensuality), and (3) it must be unambiguously
identied and ubiquitously accessible over the Internet (identiability).
2.1.3 Properties and Characteristics
e key characteristics of ontology are ease of use, comprehensibility, good forma-
tion, utility, limited proliferation, and reliance on technology (Kavi and Sergei,
1995). More particularly, it should include ease of representation and use and
also support conversion of content from one ontology to another. It must also be
easy to browse and present. Ontology should completely describe the intended
content and be internally consistent in structure, naming, and content based on
well-developed guidelines. It must ultimately aid language processing in resolving
a variety of ambiguities and making necessary inferences. Situated development
limits the size of an ontology, although presumably any piece of knowledge could
be useful. An ontology is not limited to its domain but is more developed in
the chosen domain. Acquisition and utilization are made more tractable by the
deployment of recent technologies such as faster machines, color graphical user
interfaces, graphical browsers and editors, on-line lexicons, corpora, other ontolo-
gies, semi-automated tools for consistency maintenance, and interfaces for lexi-
cographer interactions.
2.2 Types of Ontologies
An ontology can be classied by the type of knowledge it conveys (Akerkar, 2009).
A generic ontology, also known as a top ontology, species general concepts dened
independently of a domain of application and can be used in dierent application
domains. Time, space, mathematics, and other components are examples of gen-
eral concepts. A domain ontology is dedicated to a particular domain that remains
generic for this domain and can be used and reused for particular tasks in the same
domain. Chemical, medical, enterprise modeling, and other uses represent domain
ontologies. An application ontology gathers knowledge dedicated to a particular
task, including more specialized knowledge of experts for the application. In general,
application ontologies are not reusable. A meta-ontology or representation ontology
species the knowledge representation principles used to dene concepts of domain
and generic ontologies; it denes, a class, a relation, and/or a function. Ontologies can
also be classied as heavyweight and lightweight based on the expressiveness of their
30 ◾ Rajendra Akerkar
contents. e parameters for such expressiveness were introduced by McGuinness
(2003) and are summarized in Table2.1.
According to Corcho et al. (2003), a lightweight ontology includes concepts,
properties that describe concepts, relationships among concepts, and concept tax-
onomies. Heavyweight ontologies are complex and include axioms and constraints.
A systematic evaluation of ontologies and related technologies may lead to a con-
sistent level of quality and thus acceptance by industry. Future eorts may also
achieve standardized benchmarks and certications.
2.3 Parameters for Building Ontologies
Because of the complexity of the task and the many demands on ontology in terms
of usability and reusability, many engineering methodologies have been developed.
As stated earlier, ontology quality is measured in terms of criteria like clarity, coher-
ence, extendibility, minimal encoding bias, and minimal ontological commitment
(Gruber, 1995; Kalfoglou, 2000). Table2.2 briey lists the criteria. To gain the
maximum benet, ontology must be shared and reused. Existing ontologies may
be combined to create new ones. is ability makes an ontology independent,
reusable, and sharable for many applications. e parameters must be considered
during ontology engineering. Many models have been utilized in designing and
Table2.1 Parameters of Expressiveness of Ontology
Controlled vocabulary List of terms
Thesaurus Relations between terms such as synonyms provided
Informal taxonomy Explicit hierarchy (generalization and specialization are
supported) but no strict inheritance; instance of a subclass
is not necessarily also an instance of a super class
Formal taxonomy Strict inheritance
Frames Frame (or class) has a number of properties inherited by
subclasses and instances
Value restrictions Property values restricted (e.g., by data type)
General logic
constraints
Values may be constrained by logical or mathematical
formulas using values from other properties
First-order logic
constraints
Very expressive ontology languages allow first order logic
constraints between terms and more detailed
relationships such as disjoint classes, disjoint coverings,
inverse relationships, part–whole relationships, etc.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.