You are on page 1of 6

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325981519

Design of heart disease diagnosis system using fuzzy logic

Conference Paper · August 2017


DOI: 10.1109/ICECDS.2017.8390044

CITATIONS READS

46 3,958

2 authors:

Tanmay Kasbe Ravi Singh Pippal


Shri Vaishnav Vidyapeeth Vishwavidyalaya Indore Chirayu University
20 PUBLICATIONS 122 CITATIONS 17 PUBLICATIONS 98 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Tanmay Kasbe on 29 January 2020.

The user has requested enhancement of the downloaded file.


International Conference on Energy, Communication, Data Analytics and Soft Computing (ICECDS-2017)

Design of Heart Disease Diagnosis System using


Fuzzy Logic

Tanmay Kasbe Ravi Singh Pippal


Department of Computer Science and Engineering Department of Computer Science and Engineering
RKDF University RKDF University
Bhopal, India Bhopal, India
tanmay.kasbe@gmail.com ravesingh@gmail.com

Abstract—In most of the cases, heart disease results in Remaining of the paper is organized as follows. Next
death. Medical diagnosis is a difficult task and most of the time section shows literature review concerning heart disease
done by experts in domain. The aim of this work is to develop a diagnosis. In section III, database design attributes with
fuzzy expert system to identify heart disease risk in the ranges for this system are discussed. Section IV reveals
patients. There are several factors to analyze the heart disease method of designing that includes designing of fuzzy expert
in the patient and it is not an easier task, which makes the system. In section V, system testing and results with
physician’s job difficult. However, the experts want an comparison of previous research is discussed. Finally,
accurate tool which considers and identifies the risk factors on section VI concludes the paper and scope for future.
the basis of provided information. This paper uses fuzzy expert
system for diagnosis of the heart disease. The proposed fuzzy
expert system consists of three major steps i.e. fuzzification,
rule base and defuzzification. Centroid technique is used for II. LITERATURE SURVEY
defuzzification. This system has 13 input parameters and 1 Several papers have been published in the diagnosis of
output parameter. MATLAB software is used as a heart disease using fuzzy expert system, data mining
development tool. The database is taken from V.A. Medical
techniques, etc. Ali-Mehdi [1] designed a fuzzy expert
Center, Long Beach and Cleveland Clinic Foundation.
system with 94% accuracy. This paper used the dataset from
Keywords—heart disease; fuzzy expert system; matlab V.A. Medical Center, Long Beach and Cleveland Clinic
software; medical diagnosis Foundation data base. They have used MATLAB software
for implementation and also used 13 input attribute with 44
rules. Baihaqi et al. [2] developed used a data mining
techniques with 81.82% accuracy. For rule base they have
I. INTRODUCTION
used C4.5, CART and RIPPER. In this paper, they have also
Now a day the use of computer technology is essential in given comparison among all three fuzzy rule based
every filed and medical diagnosis area is not an exception. techniques and also implemented their system on 411 data
The diagnostic decision depends upon experience, expertise sets. Paul et al. [3] implemented genetic algorithm based
and use of proper method with powerful logical reasoning fuzzy decision support system for the diagnosis of heart
ability. We know very well that these fields, in which the disease. In this paper, they have used FDSS based on the
computers are used, have very high complexity and preprocessed datasets and the selected effective attributes
irregularity and the use of expert systems such as fuzzy generated by correlation coefficient by the help of required
logic, artificial network and genetic algorithm have been and WLS method. In the year of 2006,
developed. Fuzzy logic is a powerful reasoning method that
can handle uncertainty of data very well. One of the most Polat et al. [4] proposed an artificial immune recognition
important causes of death worldwide is heart attack. system and fuzzy weighted pre-processing system for
Therefore, diagnosis of the heart disease is important diagnosis of heart disease with high percentage of accuracy.
requirement in daily life but due to lots of uncertainty and In this paper, results of the system were displayed by using
risk factor sometimes heart disease diagnosis is hard for a confusion matrix. Uguz et al. [5] implemented a fuzzy
experts. When a heart attack is identified, the speed of discrete hidden model. In this paper, they have mentioned
detection is highly essential to save the life of heart attack that the performance of the Surgeon integral based method
patient and to prevent heart damage. According to the is better than the performances of artificial neural network
World Health Organization, 12 million deaths occur each (ANN) and HMM based classification systems.
year due to heart attack diseases. Khatibi-Montazerac [6] proposed a hybrid interface
The dataset which is used in this work has been taken engine to find out risk level and achieved 91.58%. Their
from UCI (Machine Learning Repository) and this dataset prediction model was based on Dempster–Shafer theory of
evidence and fuzzy sets theory. Muthukaruppan [7]

978-1-5386-1887-5/17/$31.00 ©2017 IEEE

3183
International Conference on Energy, Communication, Data Analytics and Soft Computing (ICECDS-2017)

presented a particle swarm optimization (PSO)-based fuzzy peak, thallium scan, ECG, Gender and Age [1][4][12][15].
expert system for the diagnosis of heart disease. They have The output field referred to diagnosis of heart disease in the
used decision tree (DT) for taking proper input attributes patient and the result shows that the patient has the heart
because dataset consist of lots of attributes. In this paper, risk or not. It consists of integer value from 0 (no presence)
they have given complete attribute information with their to 4 (distinguish presence (values 1, 2, 3 and 4). Here,
ranges and remarks. Oad et al. [8] developed fuzzy expert increasing value shows increasing heart disease risk.
system for detection of heart disease and also used a data
mining techniques to reduce the number of attributes. They
have used only 6 input attributes with member function IV. PROPOSED DIAGNOSIS SYSTEM
declaration in broad way. Sikchi-Sikchi [9] implemented
fuzzy expert system in MATLAB software. They have This section shows the fuzzy membership functions,
provided graphical user interface with the use of visual which have been used in the implementation part, fuzzy
studio and combination of MATLAB. expert system designing (ranges for all input attributes) and
fuzzy rule base.
Rathod-Gawande [10] proposed a fuzzy expert system
for diagnosis with accuracy of 94%. This paper consists of A. Fuzzy Membership Function
11 input attributes and 1 output attribute. This system used This work uses two membership functions in the
mamdani interface mechanism. Sijimol et al. [11] designed implementation part in MATLAB tool [11].
a fuzzy petri net system for calculation of the risk level of
heart disease. They have used V.A. medical center database Triangular function: It is defined by a lower limit a, an
and used 11 input parameter amongst 76. Moreover, they Upper limit b and a value m, where
have also calculated the output from 0-4 parameter. a<m<b. Short form of this function is
Sudhakar-Manimekalai [12] introduced a novel trimf in MATLAB tool [11].
methodology for heart disease diagnosis. They have used a
unique approach and calculate the risk factor with the
combination of Fuzzy logic toolbox and SQL. In this paper,
defuzzification is done by centroid technique. Kumar [13]
designs an adaptive neuro-fuzzy inference system for heart
disease with the accuracy of 92.00%. The author has used
13 input attributes and 1 output attribute. Opeyemi-Justice
[14] proposed an adaptative Neuro-Fuzzy Inference System
(ANFIS) which is the part of hybrid neuro-fuzzy inference
expert systems. In this paper, they have used hybrid learning
algorithm with MATLAB software.
Pathania-Ritika [15] implements a fuzzy controller for Trapezoidal function: In this function ‘a’ can be used as a
diagnosis. They used 6 Input parameter and 1 output lower limit and ‘d’ is an upper limit
attribute for identify risk level. Kasbe-Pippal [16] presented with a lower support limit b and an
a survey paper on dengue fever. They have explained all the upper Support limit c, where a < b
symptom with previous research on dengue and also < c < d. Short form of this function
mentioned the first vaccine of dengue fever. Sayad- is trangf in MATLAB tool [11].
Halkarnikar [17] presented a neural network approach. This
paper consists of Multi-Layer Perception Neural Network
(MLPNN) and also used back-propagation algorithm.
Kumar-Kaur [18] developed a fuzzy logic system for
diagnosis from the database of Parvati Devi hospital, Ranjit
Avenue and EMC hospital Amritsar and International Lab.
Sikchi et al. [19] introduced a mathematical model to
predict a risk factor with fuzzy logic. This paper explains in
detail about architecture and hierarchy of fuzzy expert
system.

III. DATABASE FOR DESIGN OF FUZZY EXPERT SYSTEM B. Fuzzy Expert System Design
The database used in this work has been taken from the The most important application of fuzzy expert system is
UCI (Machine Learning Repository) and this dataset an uncertainty problem. First step in fuzzy expert system
consists of 4 databases (Implemented on 920 Patient) from designing is determination of input and output variables.
Cleveland Clinic Foundation, Hungarian institute of The system consists of 10 input fields and 1 output field.
Cardiology, Budapest V. A. Medical Center, Long Beach, The output field calculates the presence of heart disease in
CA and University Hospital, Zurich, Switzerland [20]. the patient [15].
This dataset consists of 76 input attributes and 1 output 1) Input variables: Followig are the input table range
attribute for result. In the proposed system, 10 input values and membership functions of variable Blood Pressure
attributes are used which are blood pressure, Cholesterol, “BP”.
blood sugar, chest pain type, maximum heart rate and old

3184
International Conference on Energy, Communication, Data Analytics and Soft Computing (ICECDS-2017)

TABLE I. BLOOD PRESSURE TABLE

Fuzzy Set Value 2) Output variables: In this paper, output variables refer
Input Field Ranges
to the disease condition. On the basis of input value
<134 Low
Systolic Blood
127-153 Medium
provided by the doctor or specialist, this variable indicates
Pressure the condition of disease. It is an integer value lies from 0 to
142-172 High
(In mm hg upper)
>154 Very High 5. Triangular member function is used for all the fuzzy set
values.
In above table and figure is the example of only one
TABLE III. RESULT TABLE
attribute “BP”. Now in next table, we are providing values
of all attribute with their ranges and fuzzy set values. Fuzzy Set Value
Output Field Ranges
Fig. 1. Membership function for blood pressure
-1 – 1 Healthy
1- 2 Low Risk
Result 2–3 Risk
3–4 High Risk
4-5 Very High Risk
Fig. 2. Membership function for disease condition

TABLE II. ALL 9 ATTRIBUTE TABLE


C. Fuzzy Data Rule Base
Fuzzy Set Value
Input Field Ranges Fuzzy rule base is one of the most important part of
fuzzy interface system and quality of the result is depends
<198 Low
188-250 Medium upon this rule base. Fuzzy rule base is the collection of
Serum cholesterol single attribute or combination of attribute with AND/OR
217-307 High
281-681 Very High operators [1]. In the implementation part, 86 rules are used
0-141 Low with the proper combination. In the rule base, the entire
Maximum Heart 111-194 Medium
152-353 High variable with the proper output fuzzy set value is used. For
Typical angina example, if value of BP, OP, CP is low then result is also
0-2 Atypical angina “Healthy” but in case if values of BP, SCHOL, CP is high
1-3 Non typical- then disease condition is very high.
Chest Pain
2-4 angina
3-5 Asymptomatic Fig. 3. Fuzzy data rule base
angina
-1 - 1 No (0)
Fasting Blood Sugar
0-2 Yes(1)
<2 Low
Old Peak 1.5-4.2 Medium
2.5> High
<0.4 Normal
Electrocardiography
0.4-1.8 ST_T abnormal
(ECG)
1.8> Hypertrophy
Normal
0-3
Thallium Scan Fixed Defect
3-6
(ThaScan) Reversible
6-7
Defect
-1-1 Female (0)
Gender 0-2 Male (1)
<38 Young
33-45 Medium
Age
40-58 Old
>52 Very Old

3185
International Conference on Energy, Communication, Data Analytics and Soft Computing (ICECDS-2017)

V. RESULT AND SYSTEM DESIGN

TABLE IV. COMPARISON OF PROPOSED SYSTEM WITH EARLIER


This work has been implemented with the following SYSTEMS
table values, which are taken from UCI machine database
[19]. Figure 4 shows the rule viewer and graphical result. Diagnosis System Accuracy

Proposed System 93.33%

Fig. 4. Input data with our result Muthukaruppan [7] 93.27%

Pathania-Ritika [15] 92%

Kumar-Kaur [17] 92%

Kumar [13] 91.83%

Khatibi-Montazerac [6] 91.58%

Opeyemi-Justice [14] 90%

Baihaqi et al. [2] 81.82%

VI. CONCLUSION AND FUTURE WORK


This paper used MATLAB tool for diagnosis of heart
diseases. To do this, fuzzy expert system is used with
membership function, fuzzy input and output variables and
fuzzy rule base. The proposed system has better
performance and accuracy compared with the previous
study to diagnosis heart disease. It is very easier expert
system in the term of usability and also patient can used it
by them. From the comparison table and previous study on
the same domain, this work achieves highest accuracy of
93.33%.

Fig. 5. Matlab rule viewer


In future, research should test for other similar system or
other data sets to test its quality to produce a higher
accuracy result. In this proposed system, one can generate
more new rule base in order to get higher accuracy.

Acknowledgment
The authors would like to thank RKDF University,
Bhopal, INDIA for providing the academic support.

References
[1] Ali Adeli and Mehdi.Neshat, “A fuzzy expert system for heart
disease diagnosis,” International Multi Conferences of Engineering
and Computer Scientist (IMCECS2010), Hong Kong, pp. 1-6,
2010.
[2] Noor Akhmad Setiawan, Igi Ardiyanto and Wiga Maulana
Baihaqi, “Rule extraction for fuzzy expert system to diagnose
coronary artery disease,” International Conference on Information
Systems and Electrical Engineering (ICITISEE), Indonesia, pp.
136-141, 2016.
[3] Animesh Kumar Paul, Pintu Chandra Shill, Md. Rafiqul Islam
Rabin and M. A. H. Akhand, “Genetic algorithm based fuzzy
decision support system for the diagnosis of heart disease,”
International Conference on Informatics, Electronics and Vision
With the help of above matlab rule viewer, the result can (ICIEV), Bangladesh, pp.145-150, 2016.
be verified easily. After testing, 93.33% accuracy has been [4] Kemal Polat, Salih Güne and Sülayman Tosun, “Diagnosis of heart
achieved by the proposed diagnosis system. The disease using artificial immune recognition system and fuzzy
experimental results are compared with earlier research in weighted pre-processing,” Pattern Recognition, vol. 39, no. 11, pp.
Table V and shows that the proposed system is more 2186-2193, 2006.
efficient and accurate as compare to other existing diagnosis [5] Harun Uguz ,Ahmet Arslan, RÕdvan Sarac and oglu Ibrahim
systems. Turkoglu, “Detection of heart valve diseases by using fuzzy
discrete hidden markov model,” Expert Systems with Applications,
vol. 34, no. 4, pp. 2799–2811, 2008.

3186
International Conference on Energy, Communication, Data Analytics and Soft Computing (ICECDS-2017)

[6] Vahid Khatibi and Gholam Ali Montazerac, “A fuzzy-evidential


hybrid inference engine for coronary heart disease risk
assessment,” Expert Systems with Applications, vol. 37, no. 12,
pp. 8536-8542, 2010.
[7] S. Muthukaruppan, “A hybrid particle swarm optimization based
fuzzy expert system for the diagnosis of coronary artery disease,”
Expert Systems with Applications, vol. 39, no. 14, pp.11657-
11665, 2012.
[8] Xu DeZhi,Pinial Khan Butt and Kantesh Kumar Oad, “A fuzzy
rule based approach to predict risk level of heart disease,” Global
Journal of Computer Science and Technology, vol. 14, no. 3, pp.
17-22, 2014.
[9] Smita S Sikchi and Sushil Sikchi, “Design of fuzzy expert system
for diagnosis of cardiac diseases,” International Journal of Medical
Science and Public Health, vol. 2, no. 1, pp. 55-61, 2013.
[10] Anuradha B. Rathod and P. S. Gawande, “Design of rule based
fuzzy expert system for diagnosis of cardiac diseases,” National
Conference on Innovative Trends in Science and Engineering, vol.
4, no. 7, pp. 313-320, 2016.
[11] Sijimol A. S., Pooja Vijayan and Soji Koshy, “Cardiovascular
disease diagnosis using fuzzy petri net,” International Journal of
Engineering Research and Technology (IJERT), vol. 3, no. 4, pp.
451-453, 2014.
[12] K. Sudhakar and M. Manimekalai, “A novel methodology for
diagnosing the heart disease using fuzzy database,” International
Journal of Research in Engineering and Technology, vol. 4, no. 10,
pp. 84-89, 2015.
[13] A.V. Senthil Kumar, “Diagnosis of heart disease using Advanced
Fuzzy resolution Mechanism,” International Journal of Science and
Applied Information Technology, vol. 2, no. 2, pp. 22-30, 2013.
[14] Obanijesu Opeyemi and Emuoyibofarhe O. Justice, “Development
of neuro-fuzzy system for early prediction of heart attack,”
International Journal of Information Technology and Computer
Science, vol. 4, no. 9, pp. 22-28, 2012.
[15] Neeru Pathania and Ritika, “Implementation of fuzzy controller for
diagnose of patient heart disease,” International Journal of
Innovative Science, Engineering and Technology, vol. 2, no. 4, pp.
694-698, 2015.
[16] Tanmay Kasbe and Ravi Singh Pippal, “Dengue Fever: State-of-
the-Art Symptoms and Diagnosis”, International Journal of
Computer Science and Engineering, Vol. 4, No. 6, pp.26–30, May
2016.
[17] A. T. Sayad and P. P. Halkarnikar, “Diagnosis of heart disease
using neural network Approach,” International Journal of
Advances in Science Engineering and Technology, vol. 2, no. 3,
pp. 88-92, 2014.
[18] Sanjeev Kumar and Gursimran Jeet Kaur, “Detection of heart
diseases using fuzzy logic,” International Journal of Engineering
Trends and Technology (IJETT), vol. 4, no. 6, pp. 2694-2699,
2013.
[19] Smita Sushil Sikchi, Sushil Sikchi and Ali M. S., “Generic medical
fuzzy expert system for diagnosis of cardiac diseases,”
International Journal of Computer Applications, vol. 66, no. 13,
pp. 35-44, 2013.
[20] https://archive.ics.uci.edu/ml/datasets/Heart+Disease.

3187

View publication stats

You might also like