International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2914

ISSN 2229-5518

DESIGN OF EXPERT SYSTEM FOR MEDICAL DIAGNOSIS US- ING FUZZY LOGIC

Manish Rana

Assistant Professor (Department of Computer Engineering) M.E Computer Enggineering (Pursuing) Thakur College of Engineering & Technology

Mumbai University.

Mumbai,India manish.rana@thakureducation.org

Dr.R.R.Sedamkar

Dean Academics, Professor & Head (Department of Computer Engineering) Thakur College of Engineering & Technology

Mumbai University.

Mumbai,India rr.sedamkar@thakureducation.org

ABSTRACT

This paper work proposes a model for the diagnosis of Hemorrage, brain tumor in human brain, cardiac disease in heart and thyroid disease in case of thyroid. System uses Fuzzy logic design: Fuzzifier, inference engine, rule base and defuzzification for the following model. The model take five input: protein, red blood cells, lymphocytes, neutrophils, eosinophils and give three output: normal, hemorrhage, brain tumor for Brain disease. While in Heart disease only one input is taken: C.P.K.M.B value giv- ing output cardiac disease or not. Similarly for thyroid disease the three input are take: T-3 value, T-4 count, Ultrasenitive Hor- mone (T.S.H) giving output thyroid disease or not.The medical diagnosis for the model, fuzzy rules are formulated and applied using MATLAB simulation.The simulation result are calculated on the bases of design model.The work carried in the paper proposes to develop a control system to enhance the efficicency to diagnose a disease related to human disease.

Keywords : Fuzzifier, Inference Engine, Rule Base, Defuzzification, Simulation Result, Fuzzy Control System etc

1 INTRODUCTION

HIS project work proposed a diagnosis of the haemor- rhage, brain tumour, heart disease and thyroid disease to show the probability of the disease. The simulation of fuzzy logic results shows the probability of the disease to occur and normal result probability. Hemorrhage and tumor occur by the abnormal increase or decrease of blood cells in the cerebrospi- nal fluid. The data range of the blood cells includes red blood cells, lymphocytes, protein, neutrophils, and eosinophils. These cells are used as input parameters for the fuzzy logic system. Cells make inputs which are to be fuzzified. On the basis of fuzzy rules fuzzified outputs are collected. The output describes which disease is probable and the chances of the normality by the change in the input parameters that are blood cells. The fuzzy surfaces are obtained and graphical re- lationship between the input parameters and output are shown. Heart disease the CPK-MB test is a cardiac marker used to assist diagnoses of an acute myocardial infarction. It measures the blood levels of two variants (isoen- zymes CKM and CKB) of the enzyme phosphocreatine kinase. Thyroid-releasing hormone (TRH) is injected into the body through a vein. This hormone is naturally secreted by the hy- pothalamus and stimulates the pituitary gland. The pituitary responds by releasing thyroid -stimulating hormone (TSH).
Large amounts of externally administered TRH can suppress the subsequent release of TSH. This amount of release- suppression is exaggerated in primary hypothyroidism, major depression, cocaine dependence, amphetamine dependence and chronic phencyclidine abuse.

1.1 Motivation

In the domain of medical diagnosis, there are numerous varia- bles that affect the decision process thereby causing the differ- ences in the opinions of the practitioners. There are many un- certain risk factors, so sometimes disease diagnosis is hard for experts. Having so many factors to analyze to diagnose the disease of a patient makes the physician’s job difficult. So an accurate tool will be of a great help for an expert to consider all these risk factors and show certain results in uncertain terms. Motivated by the need of such an important tool, the professional form of artificial intelligence that is based on fuzzy logic for the diagnosis of the diseases pertaining to liver as well as other organs have been developed.

1.2 Contribution

The major contribution in this project is to develop a simulated

IJSER © 2013 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2915

ISSN 2229-5518

model. A disease is usually characterized by directly observa- ble symptoms that prompt the patient to visit a physician. A series of clinical observations are undertaken to detect the
presence of a disease. In this project “Design of Expert System for Medical Diagnosis Using Fuzzy Logic”, the first step is determination of input and output variables. There are Five input variables and Three output variable for Brain Disease, one input and two output for Heart disease and three input and 2 output for Thyroid disease. The symptoms of the dis- ease are expressed by the deviation of the observations from their normal values. All input variables their description value, associated fuzzy sets and their membership function ranges are presented.

1.3 Related Survey

The studies reported on fuzzy expert systems in medical diag- nosis covers wide spectrum including the need, importance, potential and approaches for designing of the expert systems for medical diagnosis applications [3], [4], [10], [13]. Computer assisted applications for patient’s diagnosis and treatment seems to be the more recent area of interest [2], [5], [18]. The Fuzzy Expert System has proved its usefulness significantly in the medical diagnosis for the quantitative analysis and qualita- tive evaluation of medical data, consequently achieving the correctness of results. The literature survey reveals that, the commercially available expert system shells are rigorously used to write the application specific rule-bases. It has been found that the frameworks are developed for generation of fuzzy expert systems with respect to specific diseases, general purpose diagnostic systems as well as for counseling of per- sonal health [7], [16], [17]. Design of expert system frame- works for medical treatment and prevention of high risks re- lated with the human health widened the scope for implemen- tation of fuzzy Concept in medical field [11]. Suitability of computer systems using fuzzy methods and computerized monitoring and medical decision making systems have been reported [6], [8], [9], [15]. The object oriented frameworks to construct FES are proposed [12], [14]. It has been notified that,
21% reported research is devoted towards the development of methodologies and models. The share of studies conducted at architectural development level International Journal of Ap- plied Engineering Research, ISSN 0973-4562 Vol.7 No.11 (2012)
© Research India Publications;
http://www.ripublication.com/ijaer.htm of fuzzy expert system shells and frameworks have been also found to be significant, i.e. 14%. The neuro-fuzzy approach has been used by many researchers and developed many fuzzy expert systems incor- porating artificial intelligence to it. The investigation reveals that, 13% studies are contributed to the development of neuro- fuzzy based expert systems.
Figure 1 shows the percentage distribution of studies reported
based on classification of articles in last two & half decades. The graph shows an exponential growth in the interests of various researchers for the development of fuzzy applications

in the medical field. The penetration of fuzzy concept in medi- cal field seems to be at par up to the year 2000. The recent span shows the exponential growth and more attention of research- ers toward the development of fuzzy based expert systems specific to medical diagnosis. This accelerated trend intensifies the demand for focus on the development of more intelligent medical fuzzy expert systems. Figure 1. Percentage distribu- tion of studies reported based on classification of articles.

1.4 Over view of fuzzy logic medical diagnosis control system

This work proposes diagnostic system using fuzzy logic con-
trol system and its design and simulation. Predefined patterns
and the behaviour of a process are mostly controlled by diag- nostic systems. The problems controlled by such systems in- volve suggestions for a certain treatment after identification. Diagnostic systems are in the form of an expert system based on rules. Sure patterns are explained by set of rules which is called as rule based expert system. Evaluation of rules is done after the collection of observed data. Identification of pattern and suggestion of problem linked with that pattern is given when the rules are logically satisfied.
These are the type of white blood cells used in the defensive and immune system of the body. Lymphocytes are in the bone marrow. These are also white blood cells. Eosinophils are the type of white blood cells and use in the immune system of the body. These are present in blood, bone marrow and in tissues. Protein enters the CSF through blood also protein transfer from brain to CSF.
A. Heart disease the CPK-MB test is a cardiac marker used to assist diagnoses of an acute myocardial infarction. It measures the blood levels of two variants (isoenzymes CKM and CKB) of the enzyme phosphocreatine kinase.

IJSER © 2013 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2916

ISSN 2229-5518

In some locations, the test has been superseded by
the troponin test. However, recently, there have been im- provements to the test that involve measuring the ratio of the CK-MB1 and CK-MB2 isoforms.
The newer test detects different isoforms of the B subunit spe- cific to the myocardium whereas the older test detected the presence of cardiac-related isoenzyme dimers.

1.4.1 Blood tests for thyroid

1. The measurement of thyroid-stimulating hor- mone (TSH) levels is often used by doctors as a screening test. Elevated TSH levels can signify an in- adequate thyroid hormone production, while sup- pressed levels can point at excessive unregulated production of hormone.

2. If TSH is abnormal, decreased levels of thyroid hor- mones T4 and T3 may be present; T4 and T3 levels may be determined with blood tests to confirm that their levels are decreased.

3. Auto antibodies may be detected in various disease states (anti-TG, anti-TPO, TSH receptor stimulating antibodies).

4. There are two cancer markers for thyroid derived cancers. Thyroglobulin (TG) for well differentiated papillary or follicular adenocarcinoma, and the rare

This algorithm is designed for the five fuzzy input variables. Table 1 shows the membership functions of the five variables: proteins, red blood cells, lymphocytes, neutrophils and eosin- ophils. The ranges of inputs variables are given in the table.

Table:1.1 Membership Functions of Input variables: Protein, RBC, Lymphocytes, Neutrophils and Eosinophils
Fig 1, Fig 2, Fig 3, Fig 4 and Fig 5 show the membership func- tion plots for each fuzzy input. The three membership func- tions, f1 [1], f1 [2] and f1 [3] are used to show the various ranges of input fuzzy variable “Protein” in a plot consisting of two regions as shown in Fig. 1.

1

medullary thyroid cancer has calcitonin as the mark- er.

0. 9

0. 8

0. 7

0. 6

Members hip Func t ion Plot

5. Very infrequently, TBG and transthyretin levels may be abnormal; these are not routinely tested.

To differentiate between different types of hypothyroidism, a specific test may be used. Thyroid-releasing hormone (TRH) is injected into the body through a vein. This hormone is natural- ly secreted by the hypothalamus and stimulates the pituitary gland. The pituitary responds by releasing thyroid - stimulating hormone (TSH). Large amounts of externally ad- ministered TRH can suppress the subsequent release of TSH. This amount of release-suppression is exaggerated in primary hypothyroidism, major depression, cocaine dependence, am- phetamine dependence and chronic phencyclidine abuse. There is a failure to suppress in the manic phase of bipolar disorder.

2.0 Design algorithm of fuzzy logic medical di- agnosis control system

0. 5

0. 4

0. 3

0. 2

0. 1

0

200 300 400 500 600 700 800 900 1000 1100 1200

Input Variable " Prot ein"

Figure 2 Plot of membership functions for input fuzzy variable protein for fuzzy logic medical diagnosis control system
The three membership functions, f2 [1], f2 [2] and f2 [3] are used to show the various ranges of input fuzzy variable “Red blood cell” in a plot consisting of two regions as shown in Fig.
2

Members hip Func tion Plot

1

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0

The work related to the proposed design model of fuzzy logic
medical diagnosis control system for the diagnosis of human disease like for hemorrhage and brain tumor for Brain disease
,Thyroid disease & Heart disease is discussed. It also gives the probability of normality of human.

2.1 Design Algorithm for Brain disease

0 5 10 15 20 25 30 35 40 45 50

Input Variable "RBC"

IJSER © 2013 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2917

ISSN 2229-5518

Figure 3 Plot of membership functions for input fuzzy varia- ble-red blood cell for fuzzy logic medical diagnosis control system
The three membership functions, f3 [1], f3 [2] and f3 [3] are used to show the various ranges of input fuzzy variable “Lymphocytes” in a plot consisting of two regions as shown in Fig. 3.

1

0.9

0.8

Members hip Func tion Plot

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0

50 55 60 65 70 75 80 85 90 95 100

Input Variable "Ly mphoc y tes "

Figure 4 Plot of membership functions for input fuzzy varia- ble- lymphocytes for fuzzy logic medical diagnosis control system
The three membership functions, f4 [1], f4 [2] and f4 [3] are
used to show the various ranges of input fuzzy
variable “Neutrophils” in a plot consisting of two regions as
shown in Fig. 4.

Membership Function Plot

1

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

Each membership function’s detail is shown in Table 2. Table:1.2 output membership Functions fuzzy Logic Medical
Diagnosis Control System

2.2 Design Algorithm for Thyroid Disease

This algorithm is designed for the three fuzzy input variables. Table 2.1 shows the membership functions of the three varia- bles: T-3, T-4 & TSH. The ranges of inputs variables are given in the table.

0

0 10 20 30 40 50 60 70 80

Input Variable "Neutrophils"

Figure 5 Plot of membership functions for input fuzzy varia- ble-neutrophils for fuzzy logic medical diagnosis control sys- tem
The three membership functions, f5 [1], f5 [2] and f5 [3] are used to show the various ranges of input fuzzy
variable “Eosinophils” in a plot consisting of two regions as
shown in Fig. 5.

Members hip Func tion Plot

Table:2.1 Membership Functions of Input variables: T-3,T-4 &

1

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0

0 5 10 15 20 25

Input Variable "Eos inophils "

TSH.

low medium high

1

0.8

0.6

0.4

0.2

Figure 6 Plot of membership functions for input fuzzy varia-
ble-eoinophils for fuzzy logic medical diagnosis control sys- tem
There are three output variables. The plot of membership function for each variable consists of three functions.

0

1 1.5 2 2.5

T3

Figure 6 Plot membership functions for input fuzzy variable T-
3 for fuzzy logic medical diagnosis control system

IJSER © 2013 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2918

ISSN 2229-5518

low medium high

1

0.8

This algorithm is designed for the three fuzzy input variables. Table 2.3 shows the membership functions of the three varia- bles: CPKMB, RR & HR. The ranges of inputs variables are
given in the table

0.6

0.4

0.2

0

60 70 80 90 100 110 120

T4

Figure 7 Plot membership functions for input fuzzy variable T-
4 for fuzzy logic medical diagnosis control system

low medium very high

1

0.8

0.6

Table: 2.3 Membership Functions of Input variables: CPKMB, RR & HR.

0.4

0.2

low high very high

1

0.8

0

0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

TSH

Figure 8 Plot membership functions for input fuzzy variable
TSH for fuzzy logic medical diagnosis control system

0.6

0.4

0.2

0

1

0.8

Normal High Risk

0 1 2 3 4 5 6 7 8 9 10 cpkmb

Figure 10 Plot membership functions for input fuzzy variable

CPKMB for fuzzy medical diagnosis control system

0.6

0.4

low high very high

1

0.2

0.8

0 0.6

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Normal

Figure 9 Plot membership functions for Output fuzzy variable
Normal for fuzzy logic medical diagnosis control system

0.4

0.2

0

40 60 80 100 120 140 160 180 200

RR

Figure 11 Plot membership functions for input fuzzy variable RR for fuzzy medical diagnosis control system

low high very high

1

0.8

0.6

0.4

Table:2.2 output membership Functions fuzzy Logic Medical
Diagnosis Control System for Thyroid

0.2

0

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

HeartRate

2.3 Design Algorithm for Heart Disease

IJSER © 2013 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2919

ISSN 2229-5518

Figure 12 Plot membership functions for input fuzzy variable HR for fuzzy medical diagnosis control system

Table:2.4output membership Functions fuzzy Logic Medical
Diagnosis Control System for Thyroid

3.0 Result and Discussion

Calculation has been made for the input values, PRO- TEIN=450, RED BLOOD CELL=3.02,LYMPHOCYTES = 64.8, NEUTROPHILS=2.54and EOSINOPHILS = 4.87.
This thyroid algorithm consists of three input variables: T-3, T-
4 & TSH . The rule base of this system is used to determine the
two output parameter value in case of thyroid : normal, Thy-
riod according to the three input value.
Calculation has been made for the the input values for thyroid
Disease
T-3=0. 9,T-4= 65 and TSH= 0.29 for Normal Patient.
T-3= 2.3, T-4 = 115 and TSH = 11 for Thyroid Patient
Table: 3.2 Comparsion between Simulation & Calculated value
Heart algorithm consists of three input variables: CPKMB, RR

50

40

2

30 231 1

3

20

40 60 80

Normal

80

60 1

40 23

20

2

0

40 60 80

Normal

30

20 1

3

10 23

2

0

40 60 80

Normal

& HR. The rule base of this system is used to determine the
two output parameter value in case of Heart disease : normal,
have Cardiac disease according to the three input value.
CPKMB=1,RR= 35 and HR = 0.2 for Normal Patient. CPMKB= 9, RR= 180 and HR = 1.8 for Cardiac Patient.

80

60 1

40 32

20

2

0

30

20 1

3

10 32

2

0

30

20 1

3

10 23

2

0

20 40 60

Hemorrhage

20 40 60

Hemorrhage

0 50 100

Abnormal

Figure 13 MATLAB Simulation Plot for Fuzzy Logic Medical
Diagnosis control system

The medical diagnosis system is based on fuzzy logic model. It is designed for diagnosis of disease in human brain, Heart & thyroid. This system consists of five input variables: protein, red blood cells, lymphocytes, neutrophils and eosinophils is case of Brain disease. The rule base of this system is used to determine the three output parameter value in ase of brain : normal, haemorhage, brain tumor, according to the five input values.
Table: 3.1 Comparsion between Simulation & Calculated value
Table: 3.3 Comparsion between Simulation & Calculated value Designing Process involve the designing phase of Front page consist of GUI, Patient Information , Brain Disease, Heart Dis- ease, Thyroid button as input buttons. Next below button de- veloped for Membership Function as input Protein, RBC, Lymphocytes, Neutrophils and Eosinophils. Each Membership function have three values MF1, MF2 & MF3. Training Fuzzy System button consist of training fuzzy system through Mamdani Model. Through Mamdani Model system is trained with 17rules. Fuzzy System button display the 17 rules graphics, and at last close button the close the graphics.

IJSER © 2013 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 4, Issue 6, June-2013 2920

ISSN 2229-5518

Figure: 14 Result Comparison shown in other Papers “Design Model of

Fuzzy Logic Medical Diagnosis Control System”

4.0 Conclusion & Future Work

Both the design model and simulation result are same. The designed system can be extended for any number of inputs. Normal, hemorrhage and the brain tumor all depend on the inputs protein, red blood cell, lymphocytes, neutrophils and eosinophils. We can define this system for any number of in- puts. As the inputs are the blood cells and the designed system use five blood cells as inputs, similarly we can define this sys- tem more than five inputs to get more efficient human diag- nose results.
The design work is being carried out to design state of the art fuzzy logic medical diagnosis control system in future using FPGA.

5.0 Acknowledgement

This research work was carried out in the various laboratories of Electronics Department and the Department of Computer Science GC University. Paper propose the further improve- ment in finding the suitable out on the bases of calculation to define the human disease in Normal, Certain or Uncertain mode.

6.0 Reference

[1] K.P. Adlassnig, G. Kolarz, W. Scheithauer, “Present State of the Medical Expert System CADIAG-2,” Methods of Information in Medicine, vol. 24, pp. 13- 20, 1985.

[2] Ahmed, A. Bigand, V. Lethuc, P.M. Allious, “Visual Acuity of Vision Tested by Fuzzy Logic: An Application in Oph- thalmology as a Step Towards a Telemedicine Pro- ject,”Information Fusion, vol. 5, pp. 217-230, 2004.

[3] Ali, V. Singh, “Potentials of Fuzzy Logic : An Approach To Handle Imprecise Data,” American Medical Informatics Asso- ciation, vol 2, no. 4, pp. 358- 361, 2010.

[4] Ameri, H. Moshtaghi, “Design and Development of Expert Sys- tem in Differential Diagnosis ofMaxillofacial Radio-lucent Lesi-

si- ons”,http://idt.mdh.se/kurser/ct3340/archives/ht08/papers RM08/21.pdf

[5] M. Z. Asghar, A. R. Khan, M. J. Asghar, “Computer Assisted Diagnoses For Red Eye(CADRE),” Int. J. of Compouter Sci- ence & Engineering, vol. 1, no. 3, pp. 163-170, 2009.

[6] A. Banerjee, A. K. Majumdar, A. Basu, “A Fuzzy Expert System Approach Using Multiple Experts For Dynamic Follo-up of Endemic Diseases,” Sadhana, India, vol. 19, no. 1, pp. 51-73,

1994.

[7] E. Binaghi, I. Gallo, C. Ghiselli, L. Levrini, K. Biondi, “An Inte- grated Fuzzy Logic And Web-Based Framework For Ac- tive Protocol Support,” International Journal of Medical In- formatics, vol. 77, pp. 256-271, 2008.

[8] K. Boegl, K. P. Adlassnig, Y. Hayashi, T. E. Rothenfluh, H. Leit- ich, “Knowledge Acquisition in the Fuzzy Knowledge Representation Framework of a Medical Consultation Sys- tem,” Artificial Intelligence in Medicine, vol. 30, pp. 1-26,

2004.

[9] A. Ciabattoni, T. Vetterlein, K. K. Adlassnig, “A Formal Logical Framework for Cadiag-2,” Studies in Health Technology & Informatics, vol. 150, pp. 648- 652, 2009.

[10] R. L. Engle, B. J. Flehinger, “Why Expert Systems for Medical Diagnosis are not being generally used: A valedictory opinion,” Bulletin of the New York Academy of Medicine, vol.

63, no. 2, pp. 193-198, 1987.

[11] I. Jung, G. Wang, “User Pattern Learning Algorithm based MDSS (Medical Decision Support System) Framework un- der Ubiquitous,” World Academy of Science, Engineering and Technology, vol. 36, pp. 184-188, 2007.

[12] H. H. Owaied, M. M. Qasem, “Developing Rule-Case- Based Shell Expert System,” Proc.of Int. MultiConf. of Engineers & Scientists, vol. 1, 2010.

[13] JCR Pereira, P. A. Tonelli, L. C. Barros, N.R.S. Ortega, “Defuzzi- fication in Medical Diagnosis,” Advances in Logic, Artificial Intelligence & Robotics, pp. 202-207, 2002.

[14] Y. Qu, T. Fu, H. Qiu, “A Fuzzy Expert System International

Journal of Applied Engineering Research, ISSN 0973-4562

Vol.7 No.11 (2012) © Research India Publications; http://www.ripublication.co m/ijaer.htm Framework Using Object Oriented Techniques,”IEEE Pacific Asia Workshop on Computational Intelligence and Industrial Applications, vol.2, pp. 474-477, 2008.

[15] R. Seising, “A History of Medical Diagnosis Using Fuzzy Rela- tions”, Fuzziness in Finland'04, pp. 1-5, 2004.

[16] C. Snae, M. Brueckner, “Personal Health Assistance Service Ex- pert System,” World Academy of Science, Engineering and Technology : A Biological & Biomedical Sciences, vol. 4, no. 2, pp. 109-112, 2008.

[17] M. G. Tsipouras, C. Voglis, D. I. Fotiadis, “ A Framework for Fuzzy Expert System Creation -Application to Cardiovas- cular Diseases,” IEEE Transactions Biomedical Engg., vol. 54, no. 11, pp. 2089-2105, 2007.

[18] T. Vetterlein, A. Ciabattoni, “On the (fuzzy) Logical Content of

CADIAG-2,” Fuzzy Sets and Systems, vol. 161, pp. 1941-

1958, 2007

[19] Ch. Schuh. Fuzzy sets and their application in medicine. Medi- cal University of Vienna, 2005.

[20] Chang-Shing Lee, Senior Member,IEEE, and Mei-Hui Wang A Fuzzy Expert System for Diabetes Decision Support Appli- cation.

IJSER © 2013 http://www.ijser.org

International Journal of Scientific & Engineering Research, Vo lume 4, Issue 6, June-2013

ISSN 2229-5518

2921

[21] Christian J. Schuh. Monitoring the fuzziness of human vi­

tal parameters. Medical university of Vienna, Austria

,2008.

IJSER lb)2013

http://www.ijserorq