Software Development for Blood Disease Expert System
Transcript of Software Development for Blood Disease Expert System
Abstract—In this paper, an expert system that evaluates the
software development for online blood disease expert system is
studied. The system asks 13 critical questions about the blood
tests and finally the system decides the type of the blood disease.
The data is obtained from Kirkuk General Hospital and the
system is tested on medical students from the Kirkuk University
School of Medicine. The results are satisfactory after it is tested
with 100 cases. The system is fully accurate after the diagnosis of
15 different diseases and the system is easy to use also its offer
time saving. The system is built by using forward chaining
methods the rules and parameters are determined due to a
predefined knowledge based and utilized in the expert system
called, Software development for blood disease Expert Systems
prepared by the software EXSYS CORVID. This expert system
can be used in determining the blood disease in patients and it
can be used in hospitals and universities by the physicians and
medical students etc.
Index Terms—Blood, disease, expert system.
I. INTRODUCTION
Blood disease is the most dangerous disease that infects
millions of people around the world. The disease that affects
the blood is called blood disease. These types of diseases
such as blood diseases, heart diseases, cancer and eye diseases
etc. needs continuous follow-up, those which are done by
computer application expert systems and online diagnosis
systems for these diseases. There are few types of artificial
intelligence that have practical application and expert systems
are the most frequently used one. Expert systems is the
computer program that uses rules and facts to diagnosis,
classification and solve problems in different fields and built
applications in medical , engineering , physics and chemistry
etc.
Medical expert system is an important expert system
because they are used to save lives of millions around the
world by prompting the service to the user and providing cost
savings because the user can use the systems at home. There
are two stages to build medical expert systems. In the first
stage; the medical background of the blood diseases is
collected by interviewing with doctors and the specialists. In
the second stage, the information that collected are represent
by rules and facts consist from IF and THEN parts, the IF part
presenting the reasons or symptoms and THEN part contain
the disease that recognized [1].
In this paper, an expert system for blood disease diagnosis
has been developed and implemented using CORVID
EXSYS expert system development tool. The system
Manuscript received April 22, 2015; revised January 5, 2016.
The authors are with the Yıldırım Beyazıt üNiversitesi, Çankırı cad.
Çiçek Sok No: 3, Ulus-Altındağ, Ankara, Turkey (e-mail: 125101418@
ybu.edu.tr, [email protected], [email protected]).
identifies number of diseases such as acute myeloid leukemia,
acute lymphocytic leukemia, chronic lymphocytic leukemia,
chronic myeloid leukemia bleeding& hemolysis, myeloid
aplasia, sidro plastic anemia, iron store adequate, anemia of
chronic disease, alpha thalassemia trait, beta thalassemia trait
and liver disease after some questions are asked about the
results of blood tests the system is diagnosis the type of
disease the patient was injured by. In the recent years, the
blood disease and other diseases such as cancer increased
clearly, so it’s necessary to build this type of systems for the
treatment of these diseases. The determination of the type of
the blood disease early is important for the doctors for the
treatment of these diseases and save the lives of patients.
These information used in software development for blood
disease expert system are acquired from Kirkuk General
Hospital and the system is tested by the Kirkuk University
School of Medicine students and Department of Hematology
in Kirkuk General Hospital and it’s gave the expected results
after it’s tested with 100 cases.
II. LITERATURE REVIEW
Medical expert system is one of the most commonly used
expert systems thanks to its quick servicing to distant places
and its cost saving quality. In the last decade, the use of the
medical expert systems has increased because of the observed
increase in the diseases that require seriousness care such as
blood disease, cancer, heart diseases etc.
Dendral was the first medical expert system developed by
the AI researcher Edward Feigenbaum and the geneticist
Joshua Lederberg beginning with 1960. This project
consisted of research on two main programs Heuristic
Dendral and Meta-Dendral [2].
MYCIN was blood disease expert system was
developed in the Stanford University at 1970 within five or
six years by Edward Shortliffe as doctoral dissertation using
lisp and more than 500 rules the system asked question to the
user and the user response as Yes/No. MYCIN has never been
used in practice because it’s difficult and not useable but it
gives high performance in the diagnosis of diseases another
weakness of MYCIN its conclude the result nearly within
twenty minutes [3].
In the field of genetics, the MOLGEN is the most famous
expert system that supply intelligence consultant to molecular
geneticist on the designing of experiments include the
treatment of DNA. For changing DNA material the geneticist
has different types of available laboratory techniques
(deletions, insertions, cuts, joins), different techniques for
defining the biological consequences of the changes, different
tools for measuring effects [4].
Heart Disease Program (HDP) is a large medical expert
system program for the treatment of heart diseases. Doctors
Software Development for Blood Disease Expert System
Ahmad Mozaffer Karim, Fatih Vehbi Çelebi, and Adnan Saher Mohammed
Lecture Notes on Software Engineering, Vol. 4, No. 3, August 2016
179doi: 10.18178/lnse.2016.4.3.246
enter case information about laboratory test; age and the
program determine the name of heart disease [5].
INTERNIS was educational project developed at the
University of Pittsburgh in the early 1970. The aim of this
system is to diagnose internal diseases, the rules that
represented in this system is extract from one expert Dr. Jack
Myers internal medicine chairman in the university of
Pittsburgh School of Medicine. One of the weaknesses of the
INTERNIS expert system is that it’s very slow and it can
conclude the consultation between19-30 minute, so it’s very
long [6]-[8].
The researcher in the field of expert systems can be noted
that many diseases have an expert system for diagnosis or
classification its types, but these systems are different in their
efficiency and the number of diseases they classify. For
example there is number of medical expert systems in the field
of blood diseases such as MYCIN expert system which
includes most type of blood disease, but it’s difficult to use as
we mentioned above. So we decided to build blood disease
expert system (BDES) the feature of this system it’s easy to
use and gives the results immediately.
III. EXPERT SYSTEM STRUCTURE
Expert system structure is shown in the Fig. 1. The main
components of expert system are knowledge base, inference
mechanism (engine) and user interface and knowledge
acquisition.
IV.
Fig. 1. Expert system structure.
Knowledge base
In this survey we aim to identify the mediating effect of
learning orientation on the relationship between leadership
style and firm performance. To test the propositions, a field
survey using questionnaires was conducted [9].
Inference engine
It functions like an engine, directs the search in different
directions in the knowledge base up to answer to the question.
An inference engine can be either simple or complicated,
depending on the structure of the knowledge base. There is
number of techniques used in inference engine forward
chaining, backward chaining and bi-direction [10].
Explanation
This mode allows the system to explain its conclusions and
its reasoning process. This ability comes from the AND/OR
trees created during the production system reasoning process
[11].
User interface
The user interface is communicated between the user and
the system, Success of any expert system depends on the
quality and the easiest to implement of user interface [12].
V. KNOWLEDGE ACQUISITION
Knowledge acquisition means getting knowledge of a
particular domain from some source, usually human, and
building it into a computer system. And it’s one of the difficult
tasks facing expert system developer. Transferring expert
expertise is not an easy task [13]. To understand the problem
of knowledge transferring, it is important to understand how
expert deals with his expertise.
The knowledge acquisition can be classified into three
categories: manual, semiautomatic and automatic.
Manual methods: In this method, the knowledge engineer
extracts expert knowledge and then codes it in a suitable
format. There are many known manual methods such as
interviewing, protocol analysis, observation; Manual
methods are slow, expensive, and sometimes inaccurate.
Therefore, there are some efforts for automating the
process possible [14].
Semiautomatic method: In this method, some of the
knowledge acquisition is acquired automatically without
some process. It allows the experts to build their own
knowledge bases with or without the assistance from the
knowledge engineers [14].
Automatic methods: In this method, the roles of both the
expert and the knowledge engineer are minimized or
even eliminated. Knowledge acquisition is made by the
plan architect analyst. The process of using computers to
extract knowledge from data is called knowledge
discovery [15].
VI. BLOOD DISEASE EXPERT SYSTEM (BDES)
A. Systems Goal
There are three main goals of (BDES) system. The first
one is that the systems diagnosis 15 diseases in high
performance. The second goal is the system offer time saving
within one minute the BDES system can conclude the
results .The last feature it’s easy to use and the last feature is
not available in the best previous blood disease expert
systems like MYCIN that is not usable and not used by the
normal user.
B. System Design
The Blood Disease Expert System (BDES) is a large
diagnostic online program covering most areas of blood
disease. The physician can enter patient information about the
history, physical examination, and laboratory tests, and then
the program generates detailed explanations of differential
diagnoses indicating the clinical data items which support
each diagnosis. BDES is rule-based expert systems used
forward chaining techniques to return the inferences from
knowledge base. This system consists of 15 rules each of
which consists of two parts. The IF part contains the results of
blood tests and THEN part contains the disease that
recognized according to the tests results. These rules
acquisitions from KIRKUK University hospital physician and
Knowledge
Engineer Knowledge
acquisition Knowledge
Bse
Explation Inference
engine
User
Interface
Expert
User
Lecture Notes on Software Engineering, Vol. 4, No. 3, August 2016
180
they are tested and used by hospıtal doctors and students of
the medicine college.
TABLE I: BDES RULES
NO Rules …
1
IF (Increase W.B.C )
AND (Presence of blast cell and B.M failure)
AND ( More than 5 year)
THEN (Acute
myeloid
leukemia)
2
IF (Increase W.B.C)
AND (Presence of blast cell and B.M failure)
AND (Least than 5 year)
THEN (Acute
lymphocytic
leukemia)
3
IF( Increase W.B.C)
AND (Absence of blast cell and B.M failure)
AND (Presence mature lymphocyte in bm
(peripheral blood) )
THEN
Chronic
lymphocytic
leukemia
4
IF (Increase W.B.C)
AND (Absence of blast cell and B.M failure)
AND (Myelo proliferative disorder)
THEN
Chronic
myloid
leukemia
5
IF (Decrease R.B.C)
AND (MCV high >98 fl)
AND (Hyper Segmented)
THEN Drugs
Cytotoxic
Agent
6
IF (Decrease R.B.C)
AND (MCV high >98 fl)
AND ( Poly Chromatic high reticulocyte
count)
THEN
Bleeding &
hemolysis
7
IF (Decrease R.B.C)
AND (MCV high >98 fl)
AND (Target cell Somatocyte)
THEN Liver
Disease
8
IF (Decrease R.B.C)
AND (MCV high >98 fl)
AND (Dysplasiacytopenia)
THEN
Myelodysplas
ia
9
IF (Decrease R.B.C)
AND (MCV high >98 fl) AND (Dimorphic)
THEN (Sedro
plastic
anemia)
10
IF (Decrease R.B.C) AND
(MCV Normal (78_98 fl) or Low (< 76 fl) )
AND (Normal or Reticulocyte account)
AND (Target Basophilic stippling) AND
(Increased HB A2)
THEN Beta
Thalassemia
Trait
11
IF (Decrease R.B.C) AND
(MCV Normal (78_98 fl) or Low (< 76 fl) )
AND (Normal or Reticulocyte account)
AND (Target Basophilic stippling)
AND (Normal HB A2)
THEN Alpha
Thalassemia
Trait
12
IF (Decrease R.B.C) AND
(MCV Normal (78_98 fl) or Low (< 76 fl) )
AND (Normal or Reticulocyte account) AND
(Dimorphic)
THEN
Sideroplastic
13
IF (Decrease R.B.C) AND
(MCV Normal (78_98 fl) or Low (< 76 fl) )
AND (Normal or Reticulocyte account) AND
(Low ferritin consideration)
THEN
Anemia of
Chronic
Disease
14
IF (Decrease R.B.C) AND
(MCV Normal (78_98 fl) or Low (< 76 fl) )
AND (Normal or Reticulocyte account) AND
(Normal or High ferritin consideration)
THEN Iron
Store
Adequate
15
IF (Decrease R.B.C) AND
(MCV Normal (78_98 fl) or Low (< 76 fl) )
AND (High Reticulocyte account)
THEN
Bleeding &
Haemolysis
BDES implemented by using EXSYS CORVID because its
powerful expert systems software and offer ability to build
online expert systems applications and it has very easy user
interface. The rules that mentioned above implemented by
asking 13 questions to the user to diagnosis 15 blood diseases:
Q1: What is the result of blood film?
Q2: What is the result of MCV?
Q3: What is the result of blood film and reticulocyte
count??
Q4: What about the reticulocyte case?
Q5: What about the reticulocyte count?
Q6: What about morphology of cell?
Q7: What about HB electro phoresis??
Q8: What about ferritin consideration??
Q9: What about age?
Q10: Case of hypochromia?
Q11: Case of the normal or low reticulocyte count?
Q12: Presence mature lymphocyte or myeloproloferative?
Q13: Presence of blast cell in peripheral blood and bone
marrow failure?
The Number of questions asked to the user different from
one case to another after the system asked its questions it’s
concludes the type of disease. For example to improve the
system efficient we implement the first rule. The blood film
increase and Presence of blast cell in P.B+B.M failure and the
age is 50 years.
As shown in Fig. 2, the first screen consists of the definition
goal of the system.
Fig. 2. First window in run time.
Fig. 3. First question.
As Shown in Fig. 3 the first question that asked to the user
is about result of blood disease film, and there is two chances
assume that the first one increase of White Blood Disease was
chosen.
Lecture Notes on Software Engineering, Vol. 4, No. 3, August 2016
181
The second question asked to the user is shown in Fig. 4 is
about Presence of blast cell in peripheral blood and bone
marrow Failure, also there is two answers the first one is
chosen.
Fig. 4. Second question.
The Third question asked to the user is shown in Fig. 5
about age of the case also the first answer is chosen according
to the first rule in the Table I: BDES Rules.
Fig. 5. Third question.
As shown in the Fig. 6 the system concludes the disease is
(Acute myeloid leukemia) and that’s expected result
according to the first rule in the Table I: BDES Rules. From
this case we can conclude that the BDES expert system offer
both the time saving and the usability these two features are
not offer in the MYCIN which is one of the most known blood
disease expert system.
Fig. 6. BDES decision screen.
VII. CONCLUSION
In this paper, an expert system that evaluates the software
development for online Blood Disease Expert System is
studied. The system asks some critical questions about the
blood tests after it answers the questions the system concludes
concerning the type of the blood disease.
The necessity of building such type of systems comes from
the remarkably increased number of patients of blood disease
and other diseases such as cancer during the recent years,
especially in Iraq due to the last decade’s wars [16]. The
determination of the type of the blood disease at an early
phase is important for the doctors for the treatment of these
diseases and to save the lives of the patients, the usable of the
system give chance to normal users to use the system without
doctor visiting.
The information’s that used in software development for
blood disease expert system are acquired from Kirkuk
General Hospital and the system are tested by the students of
Kirkuk University School of Medicine and Department of
Hematology in Kirkuk General Hospital and gave the
expected results.
The rules and parameters are determined by a predefined
knowledge base and utilized in the expert system called,
Software development for blood disease expert system
prepared by the software EXSYS CORVID. This expert
system may be used in determining the blood disease and can
be used in hospitals and universities by the doctors and
medical students. BDES superiority to the previous systems in
the usability, time saving and the number of diseases that
diagnosis by it with high performance. BDES can conclude
the result in one minute and easy to use, in the other hand, the
MYCIN cannot use by the normal user in real world because
it’s difficulty and it’s conclude the result within 20 minutes.
In addition to this study, the questions may be more
detailed and increase the number of diseases to develop the
abilities of the expert system.
ACKNOWLEDGMENT
I also place on record, my sence gratitude on and all, who
directly or indirectly, have lemt their hand in this venture.
REFERENCES
[1] A. Z. A. O. Samy and S. Abu Naser, “An expert system for diagnosing
eye diseases using clips,” Theor. Appl. Inf. Technol., pp. 923–930,
2008.
[2] E. A. Feigenbaum and B. G. Buchanan, “Dendral and meta-dendral:
Roots of knowledge systems and expert system applications,” Artif.
Intell., vol. 59, no. 1–2, pp. 233–240, 1993.
[3] B. G. Buchanan, E. H. Shortliffe et al., Rule-Based Expert Systems, vol.
3, 1984.
[4] E. A. Feigenbaum, “Expert systems in the 1980s,” Crit. Asian Stud.,
vol. 18, no. 3, pp. 16–28, 1986.
[5] A. Al-Ajlan, Medical Expert Systems HDP and PUFF, pp. 1–12, 2007.
[6] J. D. Myers, “The background of INTERNIST I and QMR,” in Proc.
ACM Conference on History of Medical Informatics, 1987, pp.
195–197.
[7] J. D. Myers, H. E. Pople, and R. A. Miller, “INTERNIST: Can artificial
intelligence help,” Clinical Decisions and Laboratory Use, University
of Minnesota Press Minneapolis, 1982.
[8] P. V. S. Jadhav and A. A. Sattikar, “Review of application of expert
systems in the medicine,” pp. 122–124, 2014.
[9] A. Morris, The Application of Expert Systems in Libraries And
Information Centres, Walter de Gruyter, 1992.
[10] F. Battal, Structural Risk Management of Disasters, Cankaya
University, Ankara , Turkey, 2007.
[11] C. S. Krishnamoorthy and S. Rajeev, Artificial Intelligence and Expert
Systems for Engineers, vol. 11, CRC press, 1996.
[12] S. Alshaban and A. K. Taher, “Building a proposed expert system
using blood testing,” J. Eng. Technol. Res., vol. 1, pp. 1–6, 2009.
[13] M. A. Al Ahmar, “Rule based expert system for selecting software
development methodology,” J. Theor. Appl. Inf. Technol., 2010.
[14] J. Kingston, “Linking knowledge acquisition with commonkads
knowledge representation,” in Proc. BCS SGES Expert Syst. 1994
Conf., 1994.
[15] J. H. Boose, “A survey of knowledge acquisition techniques and tools,”
Knowl. Acquis., vol. 1, no. 1, pp. 3–37, 1989.
[16] J. Binns, F. Bloom, J. Bunker et al., “Gulf war illness and the health of
Gulf war veterans: Research update and recommendations, 2009-2013
updated scientific findings and recommendations,” pp. 1–123, 2014.
Lecture Notes on Software Engineering, Vol. 4, No. 3, August 2016
182
Ahmad Mozaffer Karim was born in Kirkuk, Iraq on
April 5, 1987. In 2009, he received a B.Sc degree in
software engineering techniques from Kirkuk College of
Technology, Kirkuk, Iraq. He obtained an M.Sc degree
in computer engineering from Çankaya University
Ankara, Turkey, in 2012. He is currently a Ph.D student
at Graduate School of Natural Sciences, Yıldırım
Beyazıt University, Ankara Turkey. Also he is a systems analysis specialist
at CBKSOFT Ankara, Turkey. His research interests include multi-objective
optimization, expert systems, knowledge engineering, and artificial neural
networks .
Fatih Vehbi Çelebi was born in Kahramanmaras,
Turkey in 1963. He received his B.Sc. and M.Sc.
degrees in electrical and electronics engineering from
Middle East Technical University (METU) and
Gaziantep University Turkey, in 1988 and 1995,
respectively. He earned his Ph.D degree in
electronics/computer engineering from Erciyes
University in 2002. His research interests include fuzzy
logic, semiconductor lasers, and artificial neural networks. He is currently a
full professor and chairman of computer engineering at Yıldırım Beyazıt
University.
Adnan Saher Mohammed was born in Kirkuk, Iraq on
October 16, 1977. In 1999, he received a B.Sc degree in
computer engineering technology from College of
Technology, Mosul, Iraq. He obtained an M.Sc degree in
communication and computer network engineering from
UITEN University, Kuala Lampur, Malasyia, in 2012.
He is currently a Ph.D student at graduate school of
natural sciences, Yıldırım Beyazıt University, Ankara
Turkey. Also he is Computer Engineer in the Engineering Dept. Salahuddin
province. His research interests include computer network, computer
algorithms, and expert systems.
Lecture Notes on Software Engineering, Vol. 4, No. 3, August 2016
183