=Paper=
{{Paper
|id=Vol-2280/paper-04
|storemode=property
|title=Application of Ontologies and Semantic Web Technologies in the Field of Medicine
|pdfUrl=https://ceur-ws.org/Vol-2280/paper-04.pdf
|volume=Vol-2280
|authors=Petrika Manika,Elda Xhumari,Ana Ktona,Aurela Demiri
|dblpUrl=https://dblp.org/rec/conf/rtacsit/ManikaXKD18
}}
==Application of Ontologies and Semantic Web Technologies in the Field of Medicine==
https://ceur-ws.org/Vol-2280/paper-04.pdf
Application of ontologies and semantic web technologies in the field
of medicine
Petrika Manika Elda Xhumari Ana Ktona Aurela Demiri
Department of Department of Department of Master of Science
Informatics Informatics Informatics in Informatics
University of University of University of University of
Tirana Tirana, Tirana Tirana
petrika.manika@fs elda.xhumari@fsh ana.ktona@fshn.ed aurela.demiri2@fs
hn.edu.al n.edu.al u.al.com hnstudent.info
ontology is an essential part of any solution to the
problems of medical terminology. In Albania the
Abstract knowledge about ontology are very little compared to
developed countries, or there is almost no knowledge
The complexity and diversity of knowledge in this area. This paper will present the role of semantic
and terminology of medical science is one of network and ontology in the field of medicine. It will
the main obstacles to successful be determined the role of information technology in
interdisciplinary studies. Important data are this area and the features that need to be taken into
difficult to find and to be selected mainly due account by the technological solutions. Knowledge in
to different formats, schemes and semantics this material can serve as a beginning for anyone
they have. Extracting automatic knowledge wishing to learn and practice the ontology and
from previous practices and past medical technologies of Web Semantics. Specifically, in this
history is also very difficult due to the study were taken the patient's medical records, about
diversity of medical systems. This paper aims who is created an ontology scheme and some simple
to present a solution to these problems by query. Ontology built here is just a small part of what
using ontologies and semantic web can be done for medical records. The field of study for
technologies. In this paper we intend to the application of ontology is supposed to be applied
emphasize that the use of ontologies and precisely in medicine to the importance of maintaining
semantic web technologies like RDF, OWL data retention in this field.
and SPARQL can provide the necessary
semantics for a variety of medical domains 2. Characteristics of Semantic Web
and, moreover, can serve as tools for building
innovative solutions technology to existing The web has become an inseparable object of our daily
problems. lives and the amount of online information is growing.
In addition to simple texts which occupy most of the
1. Introduction multimedia information like graphics, audio or video
have become a prevalent part. The question is how to
One of the most challenging problems in the field of find useful information in this vast of information? It
health care is to ensure interaction between health care means to find what you need without much research
systems "(Bicer, Laleci, Dogac, & Kabak, 2005). while there is infinite information about what you are
Ontology improves interaction in healthcare systems. looking for. Traditional search engines will come to the
In this area, interaction between systems is more limits of their powers when it comes to understanding
delayed than in other areas such as finance where the content of the information. Here comes Semantic
ontology is better. Ontology can help build robust and Web Help [Bern98]. Tim Berners-Lee identifies two
more interoperable information systems. They can main objectives the Web has to fulfill [Bern97].
support the need for a healthcare process for re-use, a) The first objective is to enable people to work
transmitting and sharing personal data of the patient. together by allowing them to share knowledge between
The use of ontology in medicine is mainly focused on them.
the representation and organization or reorganization b) The second objective is to include tools that can help
of medical terminologies. If it is understandable, people analyze and manage the information they share
in a meaningful way. This vision has become known
as the Semantic Web (SW). So, initially, the term
"Semantic Web" has been invented by Sir WWW Sir
Tim Berners-Lee and has also inspired many
researches in this field. Ever since its inception, the
development of Semantic Web technologies has been
closely linked to WWW. The semantic web is an
extension of the traditional web, in the sense that the
information contained in the natural text is
supplemented by clear semantics based on the formal
representation of knowledge. It has also been
conceived as an extension of the WWW that allows
computers to search, combine, and intelligently
process web content, based on the understanding that
this content has on people. In the absence of artificial
intelligence at the human level, this can only be
achieved if the intended meaning of web resources is
clearly defined in a format that can be processed by
computers. In order to achieve this, it is not enough to Figure 1: Semantic Web Stack
store data in a machine-readable syntax, but it is
required that these data are equipped with a semantics 2.1. Why is semantic needed?
that clearly specifies the conclusions that should be
Natural language is amazing; it is hard to imagine a
drawn from the available information on our disposal.
better API for knowledge [Seg,Eva,Tay12]. Without
Often it is difficult for people to agree on the validity
much effort, we can ask a foreigner how to find a place
of the content of a document, and it is therefore much
where we want to go; we can share our knowledge of
more difficult to formalize that document in order to
different things in our community of friends; we can
make sense even for the computer. So this is an
go to the library, take a book, and learn from an author
impossible attempt. The purpose of the Semantic Web
who lived hundreds of years ago. Semantics is the
is to enable cars to have more information that so far
process of communicating enough meaning to result in
has required human time and attention. As a
an action. The sequence of symbols can be used to
consequence, the Web Semantics does not refer to a
communicate meaning, and this communication can
specific extension of WWW, but is rather an ideal
affect behavior. For example, when we read this
against which the network evolves over time.
material, what we are doing is integrating the ideas
Meanwhile any progress in this area can also be useful
expressed in these sentences with everything we knew
in applications that are not closely related to the web.
before. If the writing semantic in the material is clear,
The Semantic Web Vision is to add a syntax web so
it should help readers to create an idea about ontology
that resources are more easily interpreted by programs
and web semantics to serve as a starting point for other
(or 'intelligent agents'). As mentioned above, the
larger jobs. If the semantics is not reached, then the
Semantic Web is a vision for the future of the web in
reader will not get much benefit from the study. So
which information is given clear meaning, making it
semantics is very important and is the main thing to be
easier for machines to process and automatically
considered during any kind of work. The foundations
integrate available information into the Web. In order
of Semantic Web technologies are data formats that
to better understand the Semantic Web, the
can be used to encode knowledge for computer-based
architecture built by Tim Berners-Lee, called
processing, although the focus is on different forms of
"Semantic web stack" [Alq16] or "Semantic Web
knowledge. We summarize three key issues that
Layer Cake" (Figure 1), comes in handy.
provide conceptual support for the Semantic Web:
Model Building, search to describe the world in an
abstract way, to allow for easier understanding of a
complex reality. Knowledge-based computing,
attempting to build reasoning machines that can make
meaningful conclusions from encrypted knowledge.
Exchange information, transmit complex source
information between computers that allow us to and classify information, and different terminologies
disseminate and link knowledge at a global scale. are used together with information. By creating
different schemes, it is possible to create a unified view
2.2. The scope of the semantic web and achieve interaction between processes that use the
Semantic Web technologies and semantics provide us information. Semantic descriptions can also be applied
with a new approach to information management and to processes presented as web services. When the
processes, the basic principle of which is the creation function of an Internet service can be described
and use of semantic metadata. Metadata is a data set semantically, then the web service can be more easily
that provides information about other data detected. When existing web services are provided
[Dav,Stu,War06]. Metadata can exist on two levels. with metadata describing their function and context,
On one hand, they can describe a document, for then new services can be automatically compiled from
example a website, or part of a document, for example a combination of existing services.
a paragraph. On the other hand, they can describe
subjects within the document, for example a person, 3. Ontologies
country, or company. In both cases the most important
thing is that metadata is semantic, so they tell us about Semantic technology has the potential to provide
solutions for many limitations, providing enhanced
the content of a document and the link to other
access to knowledge based on the use of metadata
documents or even about a subject within the
processed by the machines. By using semantics we can
document. This contrast with metadata on today's Web improve the way information is presented. Basically,
site simply describes the format in which information in all semantic Web applications is the use of ontology.
should appear: Using HTML, it can be specified that a They facilitate the sharing of knowledge and reuse
given string should be displayed in bold, red color, but between agents, whether human or artificial. The word
can not specify that the string indicates a product price, "Ontology" comes from onto-logos and is the science
a copyright name, and so on. There are a number of of being. In computer science, ontology is a formal
additional services that metadata can do (Davies et al., representation of knowledge from a variety of concepts
2003). Initially, based on the meaning of information within a domain and the relationship between
we can arrange it and find it, not just the text. Using concepts. The most widespread in computing is the
semantics the systems can understand where words or definition of Gruber [Gru05]. Gruber initially defined
expressions are equivalent. When we search for ontology as an "explicit specification of a
conceptualization". From the beginning there were
'George W Bush' we can get a document that is equally
some discussions about this definition and is thought
valid referring to the "United States President".
to be incomplete. Therefore, other authors over the
Inversely systems may differ when the same word is years have improved this interpretation by making it
used with different meanings. If we search for more expressive and comprehensible. In 1997, Borst
references to the word 'Jaguar' in the context of the defined ontology as a "formal specification of a
motor industry, the system may ignore references to common conceptualization" [Bor97]. Several years
large cats or the operating system of the same name. later, Studer et al. expanded this definition by
The way information is presented can be improved by introducing a new and more complete definition:
using semantics. Instead of a search that provides a Ontology is a formal and explicit specification of
linear list of results, results can be summed up by separate conceptualization" [Stu, Ben, Fen98]. The
meaning. So a search for 'Jaguar' can provide formal term has to do with what the ontology should
documents assembled on whether they are related to be readable by the machine. Explicit means that the
cars, large cats, operating systems, or different themes type of concepts used and the limitations during their
use are explicitly defined, so clearly stated and in
all together. However, we can go further by using
details that leave no room for confusion. The meaning
semantics to merge information from all related of the concepts should be clearly defined. While
documents, removing surpluses and summing up conceptualization is an abstract, simplified picture of
where appropriate. Links between the main entities in the world we want to present [Gen, Nil87].
the documents can be represented visually. The reason Conceptualization describes knowledge on the domain
behind all this is the ability to justify and draw (domain). The divided term emphasizes that ontology
conclusions from existing knowledge in order to create should include knowledge received from different
new knowledge. The use of semantic metadata is also groups rather than individual ones.
crucial for integrating information from heterogeneous
sources, whether within an organization or across
organizations. Various schemes are used to describe
3.1. The scope for the use of ontologies • Ontology has the ability to support the integration of
knowledge and data, which can be considered as the
Ontology is developed and defined to share knowledge
most important benefit they can bring to healthcare
among researchers working in the same field. Some of
systems.
the main purposes of using ontology are: sharing
knowledge on the same domain, reuse of previously
used ontology, sharing domain knowledge from
operational knowledge. Some of the design criteria for
ontology are: clarity, coherence and extension.
Ontology consists of a number of different 4.1. Current work on medical ontologies
components. The names of these components vary There are some ontology and systems that have been
depending on the ontological language used and the created nowadays in the field of medicine. Some of
authors. But the core components are common among them are: CO-ODE: Cooperative Open Environment
different ontology. These components can be divided Development Project, Medical Informatics Group at
into two types: those that describe domain entities - the University of Manchester. LinKBase: is a
called concepts, individuals, and relationships; and knowledge base of over one million language
those that either enable the use of ontology or describe independent medical concepts. Contains an ontology
ontology itself. The common components of ontology with a formal conceptual description of the medical
include: Classes, Can be real-world concrete objects or field. MedO - a bio-medical ontology developed at the
abstract concepts. Individuals, instances or objects, Institute of Formal Ontology and Medical Information
represent the basic or atomic level of ontology. Systems, Germany. The Basic Model of Anatomy - an
Attributes, aspects, properties, features, or parameters ontology which represents a coherent body of explicit
that objects might have. Links or Relationships: Define declarative knowledge about human anatomy. The
the ways in which classes and individuals can connect Consortium of Ontology of Geneva which aims to
to one another. produce a controlled dictionary that can be applied to
all organisms as knowledge even if the role of genes
4. Ontology in the field of medicine and proteins in the cell is accumulating or changing.
The use of ontology in medicine mainly focuses on the 4.1.1. GALEN
representation and organization of medical
terminologies. Doctors have developed their GALEN and "Galen-Core" is top level ontology for
specialized languages in order to help them maintain medicine [Noy, McG]. GALEN is the abbreviation for
and communicate effectively general medical Generalized Architecture for Languages,
knowledge and patient information. Such Encyclopedia and Nomenclature in Medicine. This is a
terminologies, which are optimized for human European project (1992-1999) developed for the reuse
processing, are characterized by a considerable amount of terminology in clinical systems. GALEN was
of knowledge not clearly expressed. Meanwhile developed from the terminology based on the
medical information systems should be able to knowledge of the electronic records system. GALEN
communicate complex and detailed medical concepts. unlike most traditional terminologies provides
This is a difficult task and requires a profound analysis terminology in order to construct terminology
of the structure and concepts of medical terminology. description blocks. GALEN developed appropriate
It can be achieved by building ontology of the medical technology for [Gua]:
field in order to present medical terminology systems.
The benefits of ontology in this area: • Allowing clinical information to be captured,
displayed, modified and appear more powerful.
• Ontology can help build stronger systems and higher
interaction of information in health care. Interaction in • Support the reuse of information to integrate medical
health care is the ability of various technological records and other clinical systems.
information systems and software applications to 4.1.2. Gene Ontology
exchange data and use this information that is
exchanged. Gene Ontology (GO) is a controlled biological
terminology that is created by a bioinformatics society.
• Ontology may support the need for a healthcare It is a relatively new ontology compared to other
process to transmit and reuse patient records. ontology but has a greater impact on the bioinformatics
community. Initially, GO addressed terminologies
from three databases: Flybase, Saccharomyces domain: Spital
Genome database, and Mouse Genome database. The range: Mjek
gene database is the central repository for genomic
mapping data that results from the Human Genome. • punon_në
Later then developed three hierarchies of terms to
domain: Mjek
describe biological processes, cellular components,
range: Spital
and molecular functions. Authors have found it
reasonable to put definitions and comments about • ka_sëmundje
genes on the ontology notes.
domain: Sëmundje
5. An ontology example for medical range: Pacient
records • ka_trajtim
The example that will be shown below is the proposal
of an ontology for medical records in the republic of domain: Mjekimi
Albania. For this purpose, Kartele.owl has been built range: Sëmundja
using Protégé. The classes that are part of this ontology • varet_nga
are listed and described below:
domain: Sëmundja
• Person: a person can be a patient or a clinically
range: Trajtim
qualified person.
For each of the Object Properties, define the domain
• Mjek: a person qualified to treat people who are ill.
and range that are the class or subclass. Then there are
• Pacient: a person with a health problem. Data Properties that describe the connection between
instances and data veils. Some built Data_properties
• Diagnoze: Determining or identifying a diseased are:
condition. Diagnosis has two subclasses that are
Diagnoze_shtrimi and Diagnoze_e_daljes. • Id
• Sëmundje: a disorder of structure or function in a domain: Person
person. The disease has subclasses: range: int
Ekzaminimi_i_pergjithshem, Histori_e_semundjes dhe
• emri
Rezultatet_e_Ekzaminimit, respectively Overall
Examination, Injury History and Examination Results domain: Person
• Trajtimi: medical care given to a sick patient. range: String
Treatment has subclasses: • gjinia
Barnat_e_marra, Dhenia_e_mjekimit,Dozat_e_marra
dhe Kohezgjatja, respectively: Barn, Housekeeping, domain: Person
Dummies and Duration range: String
• Spitali: the institution providing medical treatment • adresa
and health care for people with health problems. A
hospital has doctors, nurses, and other staff that we do domain: Person
not need for our ontology. range: Double
In addition to class-subclass links that are easily • specializim
distinguished from the above figure, Object Properties
domain: Mjek
that are created through Protégé are
range: String
• ka_pacientë
Also for Data Properties we define the domain and
domain: Mjek range, but in this case the domain is a class whereas the
range: Pacient range is the type of value.
• ka_mjek
These were some Data Properties created for the Ontology defines a dictionary for researchers who
Person class, and are automatically applied to the want to share information in a field, it includes
subclasses of this class: Mjeku and Pacient. definitions that can be interpreted by computers,
definitions of field-based concepts, and links between
Only specializim (specialization) will be special for the them.
Mjeku subclass and will not apply to the Pacient.
The main reasons and advantages why one might want
to develop an ontology are:
• Sharing knowledge on the same domain.
• To reuse previously built ontologies
• To share domain knowledge from operational
knowledge
7. REFERENCES
[Bern98] T. Berners-lee. Semantic Web Road Map,
Oct 1998
[Bern97] Tim Berners-Lee. Realising the Full
Potential of the Web, Dec 1997
[Alq16] A. Alqazzaz. State of the Art of Semantic
Web. International Conference on Industrial
Engineering and Operations Management Detroit,
Michigan, USA, September 23-25, 2016
[Seg,Eva,Tay12] Tobby Segaran, Colin Evans, Jamie
Taylor. Programming the semantic web, 2012
[Dav,Stu,War06] John Davies, Rudi Studer, Paul
Warren. Semantic Web Technologies - Trends and
Research in Ontology-based Systems, 2006
Figure 1. Ontology graph
[Gru05] Gruber, Thomas R. "A translation approach
The figure shows a graph built with OWLViz where it to portable ontology specifications". Knowledge
is clearly seen that the subclasses are linked to their Acquisition, 2005
superclasses according to the IS-A connection.
[Bor97] W. Borst. Construction of Engineering
OWLViz is one of the important functionalities of
Ontologies. PhD thesis, Institute for Telematica and
Protégé, explained in the Technology Section used.
Information Technology, University of Twente,
Enschede, The Netherlands, 1997
6. CONCLUSIONS
[Stu, Ben, Fen98] R. Studer, R. Benjamins, and D.
The origin of the challenge of extracting or storing
Fensel. Knowledge engineering: Principles and
information dates back to the ancient times of
methods. Data & Knowledge Engineering, 1998
humanity.
[Gen, Nil87] M. R. Genesereth and N. J. Nilsson.
Using Web Semantic information can be exchanged
Logical Foundations of Artificial Intelligence.
more easily through different systems and they can
Morgan Kaufmann, Los Altos, CA, 1987
communicate more efficiently with each other.
Ontology is the best way to structure and model [Noy, McG] Ontology Development 101: A Guide to
information. Creating Your First Ontology, Natalya F. Noy and
Deborah L. McGuinness, Stanford University,
In recent years, the development of ontologies has
Stanford, CA, 94305
taken a tremendous shift from the product of artificial
intelligence labs to experts from various fields.
[Gua] Formal Ontology and Information Systems,
Nicola Guarino, National Research Council,
LADSEB-CNR, Corso Stati Uniti 4, I-35127 Padova,
Italy