Upload
ngonhi
View
221
Download
0
Embed Size (px)
Citation preview
An Application of Expert System for Diagnosing
Fever Caused by Viral Infection
Oktoria and Cheng-Hong Yang Department of Electronic Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung, Taiwan
Email: [email protected], [email protected]
Li-Yeh Chuang
Department of Chemical Engineering, I-Shou University, Kaohsiung, Taiwan
Email: [email protected]
Abstract—A fever, also known as a high temperature, is not
caused by an illness. It’s usually a symptom of an
underlying condition, most often an infection. In children
with fever, accompanying symptoms such as lethargy,
fussiness, poor appetite, sore throat, ear pain, vomiting and
diarrhea. Some parents think they should not go to the
hospital when children get fever, because they will be
ridiculed. As a first-aid solution, An Application of expert
system (AExS) can help parents to identify fever caused by
viral infection. It implements forward chaining as the
technique of searching through rule-base system. Some
programming code were also written in PHP for making
deduction of new fact from rules in the knowledge base. The
web based expert system enables user to diagnose the
children’s disease anytime and anywhere, just by accessing
the internet. It is very interesting and has created
considerable importance system to diagnose fever caused by
viral infection.
Index Terms—fever, symptom, expert system, forward
chaining, rule-base system, knowledge base
I. INTRODUCTION
Fever is usually associated with physical discomfort.
Many experts believe that fever is a natural bodily
defense against infection. There are also many non-
infectious causes of fever. As fevers range to 104 F (40C)
and above, however, there can be unwanted consequence,
particularly for children. These can include delirium and
convulsions. Although a fever is easy to measure,
determining its cause can be hard. Besides a physical
exam, the doctor will ask about symptoms, conditions
and medications.
The part of the brain called the hypothalamus controls
body temperature. The hypothalamus increases the
body’s temperature as a way to fight the infection.
However, many conditions other than infectious may
cause a fever. Cause of fever include bacterial infections,
viral infections, medications, illicit drugs, illness related
to heat exposure and allergies
Computer-based methods are increasingly used to
improve the quality of medical services. Artificial
Manuscript received January 1, 2016; revised June 24, 2016.
Intelligence (AI) is the area of computer science and it is
applied for decision making, diagnosis, pattern
recognition, and analysis of evidence. Furthermore, an
application of the scientific study of AI, expert systems
area able to use human knowledge through an
interference engine to solve problem that require a human
expertise [1]. This system (AExS) allows the user
(parents) to find out symptoms by providing questions
related to conditions that appeared similar to the
symptoms. Then, it diagnoses fever caused by viral
infection based on its artificial knowledge. The data
which is used by AExS came from books and a few
electronic website.
II. METHODS
Factual knowledge relevant to the problems is realized
as data or as a database or bases. The skills needed to use
the knowledge are available to solve problems, and as a
guide to acquire new facts and to learn new skills, are
realized as rules. Rule-based systems require that the
expert’s knowledge and thinking patterns should be
explicitly specified [2]. Hence, obtaining this knowledge
and writing proper rules is called the knowledge
acquisition phase. The writer is going to represent a
knowledge base of skills as a set of rules.
An expert system rule may be formulated as
if A then B (1)
where A is a set of conditions on data and B is a set of
instructions to be carried out if the rule is fired. In
forward chaining (Fig. 1), that sequence is followed.
Figure 1. Knowledge base of forward chaining
The rules are examined to see which rules are made
fireable by the data, that is, A is satisfied, and a rule or
Journal of Life Sciences and Technologies Vol. 4, No. 1, June 2016
17© 2016 Journal of Life Sciences and Technologiesdoi: 10.18178/jolst.4.1.17-21
rules selected for firing. When the rule is fired, the set of
instructions B is executed. This is the way most rule-
based expert systems work, including AExS. Forward
chaining is reasoning from facts (symptoms) to the
conclusion (fever) resulting from those facts.
The most popular type of expert system today is the
rule-based system [3]. A rule-based system consist of IF-
THEN rules, facts, and an interpreter that control which
rules is invoked depending on the facts in working
memory. It is appropriate methodology for all medical
domains and tasks for the following reasons: cognitive
adequateness, explicit experience and subjective
knowledge, automatic acquisition of subjective
knowledge, and system integration [4]. Rule based
system presents an essential technology of building
intelligent Clinical Support System for medical diagnoses
that can aid significantly in improving the decision
making of the physicians.
Computer assisted application for patient’s diagnosis
and treatment seems to be more recent area of interest.
Diagnosis is the identification of the nature and cause of a
certain phenomenon. Diagnosis is used in many different
disciplines with variations in the use of logic, analytics,
and experience to determine “cause and effect”. The
expert diagnosis (or diagnosis by expert system) is based
on experience with the system. In addition, by using this
experience, a mapping is built that associates the
observations to the corresponding diagnoses effectively
and efficiently. The experience can be provided by
human operators who translate human knowledge into a
computer language.
Figure 2. Working of medical expert
Novaliendry [5] designed Expert System to diagnose
vitamin deficiency in humans in the form of desktop-
based application. This application uses forward chaining
method and technique of depth first search engines.
Santosh [6] introduced a Diagnosis Expert System (DExS)
with forward reasoning. DExS helps to diagnose diseases
based on user’s answers to specific questions which are
asked by the system. Computer Assisted Diagnoses for
Red Eye (CADRE) is a rule-base Expert System that
assists in red eye diagnosis and treatment. It was
introduced by Asghar [7]. CADRE evaluates the risk
factors of 20 eye diseases and works just like an
ophthalmologist.
A comprehensive study of medical expert systems for
diagnosing of various diseases was presented by Jimmy
[8]. It provides a brief overview of medical diagnostic
expert systems and presents an analysis of already
existing studies. In the Fig. 2 the simulation of medical
expert systems is presented. S1 and D1 denote the first
symptom and first disease respectively. In general Si Dj
denotes the “i” symptom of “j” disease.
The logic of AExS like this: if it has a positive answer
to the symptom, it goes on with the symptoms from that
disease. If only one symptom from the disease is negative,
it moves to the first symptom from the next disease.
III. ANALYSIS AND SYSTEM DESIGN
Distinct diagnosis tools have been used as a approach
before designing the Application of Expert System
(AExS). In [8] explained that the performance of expert
system depends on these factors: (1) The accuracy and
other parameters of expert system depend on the
knowledge base. (2) The knowledge base should have
relevant knowledge. (3) There should be stress on
knowledge acquisition, a stage in which knowledge is
gathered.
The knowledge in expert systems may be either
expertise, or knowledge that is generally available from
books, magazine, and knowledgeable persons [3]. So,
symptom information for conditions involving a fever
caused by viral infection has been gathered from various
sources [9]-[11]. It may vary on each child and only the
doctor can give sufficient diagnosis.
Rule-based experts system has been used to construct
the AExS. It is better-suited representations for
explanation as their inferences can provide transparency
because of the explicit way in which the knowledge is
represented [12]. Fig. 3 shows the most important
modules that make up a rule-based expert system.
Figure 3. Structure of AExS
An end user (parent and doctor) would communicate
with the system via a user interface and explanation
facility that would interact with an expert system
inference engine. Parents only choose symptoms which
are asked by the system, while doctors can add and edit
the disease and the rule.
One of the most popular approaches to knowledge
presentation is to use production rules, sometimes called
IF-Then rules (Table I). Some of benefits of IF-THEN
rules are that they are modular, each defining a relatively
small and, at least in principle, independent piece of
knowledge. New rules may be added and old ones deleted
usually independently of other rules.
Journal of Life Sciences and Technologies Vol. 4, No. 1, June 2016
18© 2016 Journal of Life Sciences and Technologies
TABLE I. PRODUCTION RULE
Rule 1: If fever
And anorexia (eating disorder)
and headache and vomiting
and muscle pain and painful swelling of one or both parotid glands
and sore ears
Then Epidemic parotitis (Mumps)
Rule 5: If fever
and anorexia (eating disorder)
and severe headache and vomiting
and malaise (uneasiness) and cough
and nausea
and sudden high fever and low blood pressure
and bleeding gums and fingers and tip of nose feel cold
and lethargy(tiredness)
and feel agitated
and mouth cyanosis
and acting fussy / fussiness Then Dengue Fever
Rule 2:
If fever and malaise (uneasiness)
and coryza (head cold,sneezing)
and conjunctivitis (red eyes) and cough
and red flat rash (macular rash) and papular rash
Then Morbili (Measles)
Rule 5:
If fever and loss / poor appetite
and headache
and vomiting and body aches
and nausea and abdominal pain
and lethargy(tiredness)
and dark urine and fatty liver / enlargement of liver
and stools are pale or clay-colored Then Hepatitis B
Rule 3:
If fever
and anorexia (eating disorder)
and headache
and malaise (feeling tired) and cough
and aching muscles and the small red dots on the face, scalp, torso, upper
arms and legs, small bumps, blisters and pustules
Then Varicella (Chickenpox)
Rule 7:
If fever
and weakness
and swollen lymph nodes (lymphadenopathy/lymphadenitis)
and weight loss and high fever
and sweats (particularly at night) and chronic diarrhea
and opportunistic infections
Then AIDS
Rule 4:
If fever
and anorexia (eating disorder) and headache
and vomiting and muscle pain
and sore throat
and nausea and abdominal pain
and hyperemic pharyngeal
and neck stiffness and pains in the arms and legs
and having a convulsion
and suffering a neurological disorder Then Poliomyelitis
Rule 8:
If fever
and vomiting and sore throat
and feel tired and mouth sore
and sores or blisters may appear in or on the mouth and on the hands, feet, and
sometimes the buttocks and skin rash
and insomnia
and sweat out
Then Hands-foot-and-mouth disease
TABLE II. DISEASES DIAGNOSED IN AEXS
No Name of Disease
1 Epidemic parotitis (Mumps)
2 Morbili (Measles)
3 Varicella (Chickenpox)
4 Poliomyelitis
5 Dengue Fever
6 Hepatitis B
7 AIDS
8 Hands-foot-and-mouth disease
A rule based system will contain global rules and facts
about knowledge domain covered. During a particular run
of the system a database of local knowledge may also
established, relating to the particular case in hand.
The application of expert system diagnoses fever
caused by viral infection involving 8 kind of diseases
usually suffered by the children (Table II). It helps
parents to acquire the required solution regarding the
unusual disorder attack their children. The expert rules
were built up on the symptom of each kind of diseases,
and they were offered using inference tree and deduced
using forward-chaining technique.
The inference tree provides a schematic view of the
inference process (similar to a decision tree). Each rule is
composed of the symptoms and a disease. In the diagram
Journal of Life Sciences and Technologies Vol. 4, No. 1, June 2016
19© 2016 Journal of Life Sciences and Technologies
below (Fig. 4) each of these is represented by a node.
Branches connect the symptoms and diseases.
Figure 4. The inference tree of AExS
The Inference tree will make easier to draw up the
knowledge base and rules of the diagnosis of fever caused
by viral infection. Inference engine directs the search
through the knowledge base. It dictates which rule to
“fire”. The AExS implements forward chaining as the
technique of searching through rule-base system. This
method is helpful in recognizing the symptoms which is
felt by the children, while parent is given the facility to
choose the existing symptoms.
Figure 5. ERD
The Entity Relationship(E-R) diagram shows the
entities for which information need to be stored and the
relationship between those entities. AExS consists of six
entities (Fig. 5) such as disease, rule, symptom, user,
tmp_analysis and tmp_disease.
IV. RESULT
System tests, such as logic tests, debugging, rule
checking were carried out by knowledge engineer to
ensure the system would work correctly.
In the expert system, rules are required to implement
the desired knowledge. User (doctors) can add and edit
the rule. Fig. 6 show the screen of designed rules.
Figure 6. Designed rules
Figure 7. Selection of general symptoms
Figure 8. Form dialogue consultation result
AExS provides diagnostic assistance especially
concerned with fever caused by viral infection. User
(parents) only choose symptoms which are asked by the
system. Fig. 7 shows the interface screen to accept the
input values of symptoms.
In a few minutes after choosing the symptoms, the
system will determine disease suffered by the children,
the detail information, and suggestion of treatment. The
Fig. 8 below is the view of the result of the consultation
form dialog.
On the form dialog consultation result will be
displayed along with the diagnosis and solution form
based on input data from the consultations.
V. CONCLUSION
AExS contains knowledge of fever caused by viral
infection involving 8 kind of diseases. It accommodates
Journal of Life Sciences and Technologies Vol. 4, No. 1, June 2016
20© 2016 Journal of Life Sciences and Technologies
to input all kind of disease. The inference chain of this
system has strong-coupling between rules in the
knowledge base. It means “if the firing of one rule is
guaranteed to make another rule fire”. Therefore, a strong
link in the inference chain that leads from valid
symptoms to a valid conclusion (disease).
The explanation component would combine with the
inference engine to provide explanation that could follow
a consultation and provide postadvice explanation-called
feedback. Not only providing feedback, this system
should be complete with a feedforward explanation, that
provides the user with a means to find out why a
questions is being asked during a consultation (i.e.,
during the symptom input stage). Then, this system
should be complete with the experimental result to know
the system performance.
The process of constructing AExS is in many ways
similar to that of designing and building any large
computer program or a suite of programs. It is very
interesting and has created considerable importance
system to diagnose fever caused by viral infection. It was
developed with web as the basis, which will ease the user
to access the application anytime and anywhere. However
user (parents) only choose symptoms which are asked by
the system. Then the system will determine disease
suffered by the children, the detail information, and
suggestion of treatment.
ACKNOWLEDGMENT
I would like to express my gratitude to all those who
gave me the possibility to complete this paper. The
Special thank goes to DIKTI and State University of
Padang, Indonesia.
REFERENCES
[1] Kumar and A. V. Senthil, Fuzzy Expert Systems for Disease
Diagnosis, Hershey PA, USA: IGI Global, 2015, ch. 1, pp. 1-2.
[2] W. Siler and J. J. Buckley, Fuzzy Expert Systems and Fuzzy Reasoning, Hoboken, New Jersey, USA: John Wiley & Sons, Inc.,
2005, ch. 1, p. 25.
[3] J. Giarratano and G. Riley, Expert Systems: Principles and Programming, Boston, Massachusetts: Thomson Learning, 2005,
ch. 1, p. 6, 34.
[4] P. P. S. Tomar and P. K. Saxena, “Architecture for midical
diagnosis using rule_based technique,” Int. J. of the Computer, the
Internet and Management, vol. 19, no. SP1, 2011.
[5] D. Novaliendy and C. H. Yang, “The expert system for diagnosing human vitamin deficiency through forward chaining method,”
presented at the 6th ICTC, Jeju Island, Korea, Oct. 28-30, 2015.
[6] K. P. Santosh, P. S. Dipti, and M. Indrajit, “An expert system for diagnosis of human diseases,” Int. J. of Computer Applications,
vol. 13, 2010. [7] M. Z. Asghar, A. R. Khan, and M. J. Asghar, “Computer assisted
diagnoses for red eye (CADRE),” Int. J. of Computer Science &
Engineering, vol. 1, no. 3, pp. 163-170, 2009. [8] S. Jimmy, G. Dinesh, and B. Abhinav, “Medical expert systems
for diagnosis of various diseases,” Int. J. of Computer Applications, vol. 93, 2014.
[9] T. H. Rampengan and I. R. Laurentz, Penyakit Infeksi Tropik Pada
Anak, Jakarta, Indonesia: EGC, 2007. [10] Symptoms of Tropical Diseases – site. [Online]. Available:
http://www.rightdiagnosis.com. Visited on 08-04-16). [11] Fact sheets: Infectious disease – site. Online]. Available:
http://www.who.int. Visited on: 25-04-16).
[12] K. Darlington, “Designing for Explanation in health care applications of expert systems,” SAGE Journals, April 28, 2011.
Oktoria is a PhD student of the Department
of Electronic Engineering at National
Kaohsiung University of Applied Sciences, Taiwan since Fall 2015. He received his
Bachelor Degree (S.Pd) from State University of Semarang, Indonesia, 2005. And his Master
Degree (M.T) was received from Gadjah
Mada University, Indones ia, 2007. Since 2008 he has been being a lecturer at State
University of Padang, Indonesia. His research
interest are in genetic algorithm, expert system and bioinformatics.
Cheng-Hong Yang is a professor of the Department of Electronic Engineering at
National Kaohsiung University of Applied Sciences, Tai-wan. He received his M.S. and
Ph.D. degrees in computer engineering from
North Dakota State University in 1988 and 1992, respectively. His principal research
interest now are in evolutionary computation, bioinformatics, and assistive tool
implementation.
Li-Yeh Chuang is a professor and director of
the Department of Chemical Engineering & Institute of Biotechnology and Chemical
Engineering at I-Shou University, Kaohsiung,
Taiwan. She received her M.S. degree from the Department of Chemistry at the
University of North Carolina in 1989 and her Ph.D. degree from the Department of
Biochemistry at North Dakota State
University in 1994. Her main area of research are Bioinformatics, Biochemistry and Genetic Engineering.
Journal of Life Sciences and Technologies Vol. 4, No. 1, June 2016
21© 2016 Journal of Life Sciences and Technologies