Vous êtes sur la page 1sur 5

JOURNAL ITSMART Vol 4. No. 1.

June 2015 ISSN: 2301-7201

EXPERT SYSTEM FOR DIAGNOSING INFECTIOUS DISEASE


USING FORWARD CHAINING

Noor Ahmad Aniq Mutsaqof Wiharto ST M.Kom Esti Suryani S.Si M.Kom
Department of Informatics, Faculty of Science, Department of Informatics, Faculty of Science, Department of Informatics, Faculty of Science,
Eleven universities in March Jl. Ir. Sutami Eleven universities in March Jl. Ir. Sutami Eleven universities in March Jl. Ir. Sutami
36 A Surakarta 57126 36 A Surakarta 57126 36A Surakarta 57126
kupersemar@gmail.com wi_harto@yahoo.com suryapalapa@yahoo.com

ABSTRACT in the abdomen), hypertension (high blood pressure), diseases of the appendix
(appendectomy) and gastritis (stomach pain) [4]. This shows that infectious diseases
Infectious diseases is one of the most important health problem in
must be quickly diagnosed and dealt with to avoid getting worse. However, the number
developing countries, including Indonesia. Infectious diseases are often
of specialist doctors in Indonesia has not been sufficient. Especially pediatricians,
suffered from diarrhea, fever typhoid, scarlet fever, pneumonia. This shows
gynecologist and obstetrics, surgery, internal medicine, anesthesia and medical
that infectious diseases must be quickly diagnosed to avoid getting worse.
rehabilitation [5]. The needs of the specialist physician in 2014 amounted to
Infectious disease is an infectious disease that is easy to attack the child,
because the child does not have a good immune system. However, the
number of general practitioners in Indonesia insufficient and uneven 29 452, whereas availability in 2013 only 20 602 specialists [6].
distribution of doctors. Therefore, it is necessary to have a system that can
help medical personnel non-physicians to diagnose infectious diseases for Another problem occurs in an uneven distribution of doctors. Doctors
the handling of infectious diseases more quickly resolved. distribution centers are in Java and Bali as well as in the province of North
Sumatra and South Sulawesi. According to the head of the Agency for Human
Resources Development and Empowerment of Health Ministry of Health in
The method is used as an inference engine is 2010, Bambang Giatno, until 2010 there were an estimated 1,600 health
centers, especially in remote areas do not have a doctor on call as a result of
Forward chaining. This method has the same logic concept with doctors
lack of equitable distribution [7]. Seeing these facts necessary to have a
when diagnosing a patient as based if-then. forward chaining is a method
that has a search strategy which initiated the process of collection of data system that can help medical personnel non-physicians to diagnose infectious
or facts, from these data produce a conclusion. diseases.

Expert systems will assist non-physician medical personnel to detect


Tests carried out by a young doctor (Co-Ass) in the District General disease and got information solutions

Hospital Sunan Kalijaga Demak. Testing was conducted on 50 patients treatment early in order to take immediate action [8]. Expert systems are
who generate 6 failures in detecting infectious diseases. Values ​obtained also expected to solve the problem is quite complicated truth can only be
accuracy was 88%. The results of the test can be concluded that the solved with the help of experts [9].
method
forward chaining can be implemented to diagnose infectious disease by The components contained in the expert system include a user
doing the stages of knowledge acquisition and knowledge representation. interface, knowledge base, knowledge acquisition, inference engine,
explanation facilities,
workplace, and improvement of knowledge [10]. The inference engine in an
Keywords :
Forward Chaining, Infectious Diseases, Expert System.
expert system is essential because it is the brain of the expert system. One of
the methods used in the inference engine of the expert system is a method forward
chaining.
1. PRELIMINARY This method has the concept of logic that is easily understood by doctors
Disease infection is a collection of types because based if-then. forward chaining is a method that has a search
an easy disease in children caused by viral infections, bacterial infections, strategy which initiated the process of collection of data or facts, and from
and parasitic infections [1]. Infectious diseases is one of the most important these data sought a conclusion [11]. forward Chaining already widely
health problem in developing countries, including Indonesia. Based on implemented in the field of medicine, for example in diseases of ENT (ear,
Household Health Survey of 2007, the leading cause of death among other nose and throat), asthma and respiratory problems, and eye diseases [12]
28.1% were caused by infectious and parasitic diseases, 18.9% were [13] [14].
caused by vascular disease, and 15.7% were caused by respiratory
diseases [2]. Another thing to consider is the number of child mortality in Method forward chaining suitable for diagnosis of the disease, being
Indonesia, more than 200 children die per 100,000 births figures. This able to identify the type of disease based on symptoms that met [15]. forward
figure is still far from the MDGs ( Millennium Development Goals) Countries chaining an inference engine that conducts a search of a problem to a
Indonesia in 2015 is 102 per 100,000 births [3]. solution, in which forward chaining an inference datadrive, inference starting
from the information available and the new conclusions [16].

Diarrheal diseases, fever typhoid, dengue fever, upper respiratory tract Based on the above background author wants to do research with the
infection (influenza, tonsillitis, laryngitis), pneumonia, and fever of unknown title "Expert System for Diagnosing Infectious Diseases Using Methods forward
cause (observation febrile) is an infectious disease that is among the 10 Chaining ".
main diseases of hospitals in Indonesia. Whereas other diseases, among The aim of this study is to produce an expert system using the method forward
others: dispepsi ( uncomfortable nuisance chaining in the diagnosis of infectious diseases.

43
JOURNAL ITSMART Vol 4. No. 1. June 2015 ISSN: 2301-7201

The benefits of this research is expected to help non-physician medical Table 2. 1 Table Kep u decision- [18]
personnel to facilitate the earliest possible diagnosis of infectious diseases. Hypothesis Hypothesis Hypothesis Hypothesis
1 2 3 4
Facts A v v v
2. BASIC THEORY
There are several theories that support in this study are as follows. fact B v v v

fact C v v

2.1 Expert system


Facts D v
An expert system is a computer-based application that is used to
solve the problem, as is thought by experts. Specialists in question here is,
fact E v v
people who have special skills that can solve problems that can not be
resolved by lay people [17]. 2.5 Decision Trees ( Decision Tree)
The decision tree is a form of knowledge representation is used to
model the problem consists of a series of decisions that leads to a solution
2.2 Expert System Architecture [19]. The decision tree is made with reference to the decision table. An
Expert systems composed by two main parts, namely the example of the decision tree contained in figure 2.2.
development environment and environmental consultancy.
The development environment is used to enter development experts into A

the expert system environment, the environment used by non-expert Yes Not

consultation to acquire knowledge and expert advice [10]. Expert system


B B

architecture can be seen in Figure 2.1. Yes Not Yes Not

end
C E C

Yes Not
Yes no Yes

end
D

H1 E H2 end
no Yes

Yes No
H4 end

end No
H1

Figure 2.2 Decision Tree [18]

Forward trace 2.6 ( Forward Chaining)


forward chaining a search strategy which initiated the process of
collection of data or facts, of the data sought a conclusion that a solution of
the problems faced. Looking kaidahkaidah inference engine in the
knowledge base that premise is in accordance with these data, then from
the kaidahkaidah obtained a conclusion. forward chaining start the search
process with the data that this strategy also called datadriven [ 10].
Expert System Architecture Figure 2.1 [10]

examples of process forward chaining shaped the rules of production [18]:


2.3 Knowledge acquisition
The acquisition of knowledge is done to gain knowledge from
IF condition 1
experts or other sources (sources documented, books, censorship, file computers,
AND condition 2
etc.) [10]. In the process of knowledge acquisition,
AND condition 3
someone engineer knowledge
THEN conclusion.
bridge between experts with a knowledge base ( knowledge base). Engineers
gain knowledge from expert knowledge, process it and put it together with
the expert knowledge base with a particular format. 2.7 Infectious diseases
Infectious diseases are a set of types of diseases are easily strike
children caused by viral infections, bacterial infections, and parasitic
2.4 table Decision infections. [1]. There is a wide - range of infectious diseases among
Table decision is a way to diarrhea, fever
documenting knowledge. Decision table is a matrix of conditions that are typhoid, dengue fever, upper respiratory tract infection
considered in the description of the rules [18]. Example of decision table (Influenza, tonsillitis, acute laryngitis), pneumonia, and observation of
are in Table 2.1. febrile [ 6].

44
JOURNAL ITSMART Vol 4. No. 1. June 2015 ISSN: 2301-7201

3. RESEARCH METHODS 4.1 Knowledge acquisition


Research methodology for completing the study authors described Based on the results of interviews conducted by the author in
the measures as follows: hospitals Sunan Kalijaga Demak with dr. Nur Budi Cahyani, Sp.A, he
suggested that the case studies used for this study is 6 infections are
included in the 10 largest illness in hospital Indonesia. Infectious diseases
Study of literature include diarrhea, fever typhoid,

scarlet fever, pneumonia, observation febrile, and ISPA (upper respiratory


Knowledge acquisition
infection). ISPA itself can be divided into three diseases namely influenza,
tonsillitis, and acute laryngitis. The case studies used in the study contained 8
to 10 types of disease symptoms, one of which is a common symptom. Fever is
Knowledge representation
• table Decision
a common symptom, while other symptoms include cough, runny nose,
abdominal pain, shortness of breath, CHAPTER (defecate) liquid is more than
• Decision trees
4 times daily, bleeding (red spots on the skin, nosebleeds, BAB bloody), tongue
• analysis forward Chaining
dirty, sick time swallowing, and snoring during sleep.

examination

Figure 3.1 Flow Research 4.2 Table Design Decisions


The design of the decision table is done by identifying rule from knowledge
3.1 Study of literature experts then enter rule into the groove for the search to be traced and
Studies literature conducted by studying translated into Table 4.1.
literature to understand the theory derived from books and articles and
journals are sourced from the internet. Literature study was conducted to
study matters relating to research in general, expert systems, Table 4.1 Ta b el Ke p utusa n peny a kit inclusive f reflection
P00 1 P00 2 P00 3 P00 4 P00 5 P00 6 P00 7 P00 8

forward chaining, various infectious diseases and also symptoms of what is


G00 1 V V V V V V V V
perceived as a result of infectious diseases.
G00 2 V V V V
3.2 Knowledge acquisition
G00 3 V V V V
acquisition phase called knowledge too stage
collection knowledge. This phase is done input Data symptoms, diseases G00 4 V V V
and medical advice obtained from Hospital Pediatrician Sunan Kalijaga
Demak, dr. Nur Budi Cahyani, Sp.A. G00 5 V

G00 6 V

3.3 Knowledge representation G00 7 V

Knowledge representation stages are stages used for coding


G00 8 V
knowledge expert system. Knowledge representation consists of the
manufacture of decision tables, decision trees and design inference G00 9 V V

analysis forward chaining


G01 0 V

Table 4.1 include symptoms column (G) and a column of disease (P).
3.4 examination There are 10 symptoms of infectious diseases, with code G001 to G010. Details
This stage testing of applications with the aim of ensuring the symptoms - symptoms of infectious diseases in Table 4.2. There are 8 types of
functionality of the application can be run the same as what was expected. diseases in the fields of disease that is encoded with the letter P001 to P008.
And also will be tested by comparing the results of the patient's disease Details of infectious diseases in Table 4.3.
have been diagnosed by a doctor with the diagnosis system. Furthermore,
the accuracy of calculation as follows
Table 4.2 Infectious Disease Symptoms
Code Symptoms name
••••••• = •••••ℎ •••• •••••••••• ••••••• × 100%
•••••ℎ ••••••••ℎ•• ••••
G001 Fever

G002 Cough
4. RESULTS AND DISCUSSION
Discussion expert system for the detection of infectious diseases G003 Cold
through several stages. Starting from the acquisition of knowledge,
Abdominal pain G004
knowledge representation, implementation and
system testing. Discussion expert system for the detection of infectious diseases G005 Hard to breathe
through several stages. Starting from knowledge acquisition, knowledge
representation, implementation and testing of the system. G006 CHAPTER liquid is more than 4 times a day

G007 Bleeding (nosebleeds, red spots on the skin or bleed


BAB)

45
JOURNAL ITSMART Vol 4. No. 1. June 2015 ISSN: 2301-7201

G008 tongue dirty (covered in white or yellow membrane attached to the Figure 4.1 Decision Tree for Infectious Diseases
surface), edge hyperthermia ( redness of the tongue edge), and tremor
G009 4.4 analysis Inference forward Chaining
Pain when swallowing analysis inference forward chaining can form a
G010 snore rule. rule It describes the disease based on symptoms. Rule consists of the
input information in the form of symptoms ( if),
Advanced symptoms ( and) and the conclusion that the form of the diagnosis of infectious
Table 4.3 Infectious Diseases diseases ( then). Rule 1 provides IF G001 AND G004 AND
Code Disease name G006 THEN P001. This means that if the first symptoms (fever), 4 symptoms (abdominal
P001 Diarrhea pain), and symptoms 6 (liquid chapter> 4 times daily) suffered by the patient, the system
P002 Fever typhoid diagnosing these patients suffer from P001 (diarrhea).
P003 Dengue fever
P004 influenza
P005 pneumonia ( pneumonia) • rule 1 • rule 5
P006 observation Febris IF G001 IF G001
P007 tonsillitis ( Tonsillitis) AND G004 AND G002
AND G006 AND G003
P008 Tonsillopharyngitis ( Sore throat)
THEN P001 AND G005
THEN P005
4.3 Design Decision Trees • rule 2 • rule 6
Figure 4.1 is based on Table 4.1. G001 is a common symptom for IF G001 IF G001
people with infectious diseases of fever. The next symptom is abdominal AND G004 THEN P006
pain. Based on the symptoms of abdominal pain can be grouped diseases AND G008
suffered by patients whether they have symptoms of abdominal pain or not. THEN P002
If the symptom questions have answers "Yes" then the decision tree will be • rule 3 • rule 7
continued questions with answers "yes" and the search will be continued so IF G001 IF G001
that it will happen decision infectious illness symptoms suffered by patients. AND G004 AND G002
AND G010 AND G003
THEN P003 AND G009
Examples of the first symptoms patients experienced symptoms of THEN P008
fever, then the system will display the question "Does the patient have
abdominal pain?" If the patient does not experience then continued with the
question "Is the patient cough?". If the patient is not experiencing system
will diagnose the patient is suffering from febrile observation. The system • rule 4 • rule 8
will stop to ask questions when they are achieving results, namely patients IF G001 IF G001
diagnosed infectious disease or the system can not detect the patient's AND G002 AND G002
illness. AND G003 AND G003
THEN P004 AND G009
Fever AND G010
(G001)
THEN P007
Yes detected

not been
Stomach
pain
There has 4.5 Testing
(G004)
Tests performed on patients of diarrhea, amounting to 12 people
only system capable of detecting a number of 11 people. Because the
Yes

Chapter Liquid>
4x a day
Not system can not detect the patient's illness. For example, in making a
(G006)
diagnosis of diarrheal diseases, the system can conclude that the patient
Yes
photo
was suffering from diarrhea when meet rule as follows:
Bleeding
Cough
Diarrhea G010
(G002)
(P001)
IF Fever
Yes
Not
Yes
AND abdominal pain
Dengue fever (G008)
Dirty No Bank
Tongue, observation Febris
(P003) Hipertremi,
Colds (P006)
AND CHAPTER liquid> 4 times daily
(G003)
Tremor
THEN Diarrhea

Yes
Yes Not
but sometimes patients do not experience or feel any of the symptoms
Not
Shortness
has not been
listed in route infectious diseases. For example, patients only experience
Typhoid of Breath
(P002) has not been
detected
(G005) detected
symptoms of fever and liquid BAB> 4 times a day, so that the system can
Yes Not not detect that a patient suffering from diarrheal diseases. Just as in fever typhoid,
influenza, tonsillitis and strep throat.
Pneumonia
Pain when
(P005)
swallowing
(G009)
Table 4.4 Ta bell Testing
Yes

Snoring during
No. Disease name Total Total
sleep (G010)
(P004) Not
Diagnosis Diagnosis
Influenza
Doctor System
Yes Not 1 Diarrhea 12 11
Sore throat 2 Fever typhoid 6 5
Tonsillitis
(P007)
(P008) 3 Dengue fever 3 3
4 influenza 10 9

46
JOURNAL ITSMART Vol 4. No. 1. June 2015 ISSN: 2301-7201

5 Pneumonia 5 5 national / 14/10/01 /


6 observation Febris 3 3 ncr7ix-idi-call-number-dokterspesialis-still-skimpy.
7 tonsil 5 4 [6] Ministry of Health, "Health Human Resource Availability and Readiness
8 Inflammation 6 4 in the Target Achievement RPJMN, "Center for Health Human Resource
Acute throat
Planning and Utilization, Surabaya, 2012. [7] Higher Education, Portrait availability
total 50 44
and Power Requirements Doctor,
Jakarta: Directorate General of Higher Education Ministry of Education
Table 4.4 shows that the testing of 50 patients resulted in six failures
in the detection of infectious diseases. The testing accuracy value is ( 44 and Culture, 2010. [8] The US Broto, "Design and Implementation of Expert
System
x 100% = 88%.
50) for the Analysis of Internal Medicine, "University of Diponegoro, Semarang
Tests carried out by a young doctor (Co-Ass) who are doing practice at the 2010.
Polyclinic Hospital Children Sunan Kalijaga Demak for 6 days from 3-8
[9] L. Handy and T. Sutikno, "Expert System for
November 2014.
ENT Diseases Diagnosis Based Web with "e2gLite Expert System Shell",
Number diagnosis between different systems and diseases due to
" Journal of Industrial Technology, vol. Vol.
the testing system can not detect the patient's illness. This is because the
XII, no. 1, pp. 19 -26, 2008.
patient does not feel or experience the symptoms that prompted by the
system. [10] E. Turban, Sharda R. and D. Delen, Decision Support
System and Business Intelligent Systems Ninth Edition, New
Jersey: Pearson Education, 2011. [11] F. Sari, "Expert System to Know
4.6 Conclusions and recommendations the Aspects
Someone Personality Tests Based on EQ (Emotional Quotien) with
4.6.1 Conclusion
Based on the research results can be concluded that by doing Forward Chaining method, "University Putra Indonesia YPTK,
phase of knowledge acquisition and knowledge representation expert Padang, 2011. [12] Suraya," Expert System to Diagnose Disease ENT
system this successful
implemented. Method forward chaining used as an inference engine in which Based on the symptoms for Determining Alternative Medicine using
the inference engine serves as the brain of the expert system. Medicinal Plants, " National Seminar on Application of Science &
Technology (SNAST) Period III,
The results obtained from the acquisition of knowledge is to use pp. 337-346, 2012.
an infectious disease that included the 10 biggest disease in a hospital in [13] AR Tohir, "Design of Expert Systems Applications for
Indonesia. Following the acquisition of knowledge is done, then Diagnosis of Asthma and Respiratory Disorders (Case Study Houses
represented then implemented an expert system testing was conducted on Skait Doctor Soetomo), "High School of Computer Science, Surabaya,
50 patients. There are 6 failure detection of 50 tests performed. Accuracy 2012. [14] R. Reisa, Jusak and P. Sudarmaningtyas," Expert Systems
value is calculated by comparing the value of success in the detection
system with the number of patients testing. Values ​obtained accuracy of
for Eye Disease Diagnosis, " Journal of Information Systems and
88%.
Computerized Accounting (JSIKA) Vol.2 No.2, pp. 30-
40, 2013.

[15] DR Fitriana, "Designing Expert System Diagnosis


Penyaikit Heart with Forward Chaining Web-Based Method,
4.6.2 Suggestion
"Muhammadiyah University Ponorogo, Ponorogo, 2003.
Suggestions that can be used for further research are:

1. need for update knowledge base so more infectious diseases that [16] R. Sofa, D. and A. Susanto Destiani, "Development
can be diagnosed Applications for Disease Diagnosis Expert System for Rice, " Garut
2. Adding another method to avoid disease Technology High School algorithms Vol. 09 No. 03, pp. 1-8, 2012.
that can not be diagnosed because the patients do not feel or
experience symptoms. Eg by using certainty factor. [17] Kusrini, Expert Systems Applications, Yogyakarta: Andi, 2008. [18] S.

Hartati and S. Iswanti, Expert system &


Development, First Edition, Yogyakarta: Graha Science 2008.
5. BIBLIOGRAPHY
[19] J. Giarrattano and G. Riley, 1994. Expert System Principles
[1] Rampengan T. and I. Laurentz, Tropical Infectious Diseases in Children, and Programming, Boston: Publishing Company, 1994.
Jakarta: EGC Medical Books, 1997. [2] SK Nasution, "Isolation of Bacteria
Producing Antimicrobial From
Soil samples Sibolangit Nature Park and ability to Inhibit Growth of
Microbial Pathogens, "University of North Sumatra, Medan, 2012. [3]
UNICEF, Assessment Summary Maternal and Child Health,

Jakarta: UNICEF Indonesia, in 2012.

[4] Ministry of Health, "Health Profile of Indonesia," DG


Efforts of Health, Jakarta, in 2009.

[5] Republika, "IMA Specialist Doctors Still Call Number


Minimal, " 1 10 2014. [On line]. Available:
http://www.republika.co.id/berita/nasional/jabodetabek-

47

Vous aimerez peut-être aussi