You are on page 1of 6

1.

Heart Disease prediction using machine learning techniques


A Survey

R.Herini, S.Kavinpreetha, G.Priyadharshini


Department of computer science _________________________________________________________________

Abstract: Heart plays the important in maintain life of the body. In this world many people died due to heart
disease. Heart related disease is also known as cardiovascular diseases (CVDs). Now a day’s technology has been
improved a lot by that we can predict heart disease using UCI repository of the patients. In this Machine learning
and data mining algorithms has plays an important role in predicting the heart disease. In this rapid growing world
machine learning plays a vital in medical industry. It is important to predict the heart related disease in a patient.
This paper presents the survey of techniques and its accuracy of machine learning algorithms used in heart disease
prediction. The machine learning uses supervised machine learning. The supervised machine learning is a trained
data. Some algorithms are K-Nearest Neighbor (K-NN), Support Vector Machines (SVM), Naïve Bayes, Decision
tree and Random Forest (RF)

Keywords: Heart Disease, Cardiovascular disease, SVM, NN, K-NN, Naïve Bayes.

______________________________________________________________________________

Introduction:

Cardiovascular disease (Heart disease) is a serious problem and also the widespread disease in the world. It is
estimated that around 18.6 million deaths caused due to the heart attacks. It occurs due to the change of life, stress,
restless, poor diet, obesity, high blood pressure, smoking and hypertensions. Some of the symptoms for heart disease
are chest pain, felling week or faint, pain in jaw or neck and shortness of breath. Once upon a time the heart disease
is the problem of old one (elders) but now it also occurs for younger people. The treatment and diagnosis of
cardiovascular disease is complex.

Now a day’s technology helps to predict the heart disease with the help report of the patients. Using algorithms
machine learning analyzes the reports. Data mining is the collection of information that is used for analyzing the
data. Machine learning uses supervised machine learning. It is the trained data. The supervised machine learning is
of two types one is regression and another is classification. For the prediction classification model is used. This
model supports the techniques such as Random Forest, Support Vector Machine, K-NN, Neural Network and
decision tree. In recent times these algorithms become more important to predict the cardiovascular diseases (Heart
related diseases).

Methodology:
Dataset:

The dataset is the collection of information (medical analytic) from UPI machine learning repository.

Table 1: Attribute and description of heart diseases dataset

Sl.no Name Code Type Value

1 Age Age Continuous Age of patient


(in years) (Numeric)
2 Sex Sex Discrete Male=0, Female=1
(in numbers) (Nominal)
3 Chest pain Cp Discrete 4 Types:
(Chest pain type in (Nominal) 1->Typical angina=1
mm Hg) 2->Atypical Angina=2
3->non-anginal pain=3
4->asymptomic=4
4 Rest blood Trestbpd Continuous 92-200
pressure (Numeric)
5 Serum Serum Cho Continuous 126-564
cholesterol (in mg/dl) (Numeric)
6 Fasting Fbs Discrete Yes=1, No=0
blood (Nominal)
sugar
7 Rest Restecg (Rest Discrete Normal=0
electrocard electrocardiographic (Nominal) Having ST-T wave abnormality = 1
iograph results) Left ventricular hypertrophy = 2
8 Max heart Thalach Continuous 82-185
rate (Maximum heart (Numeric)
rate achieved)
9 Exercise Exang Discrete Yes=1,No=0
induced (Nominal)
angina
10 ST oldpeak ST Continuous 71-202
depression (ST depression (Numeric)
induced by exercise)
11 Slope peakslope Continuous 1-3
(Numeric)
12 No. of Ca Continuous 0-3
vessels (number of (Numeric)
vessels(0-3)
coloured by
fluoroscopy)
13 Thalassemi Thal Discrete Types:
a (The defect type of (Nominal) 1->normal=3
heart) 2->fixed defect=6
3->reversible defect=7
14 Num(class Class Discrete Status:
attribute) (identification of a (Nominal) 0-nil risk
heart attack) 1-low risk
2-potential risk
3-high risk
4-very high risk

Survey on techniques and accuracy of heart disease prediction

S.No Month and Authors Title Techniques Accuracy


Year of
publication
1 Sep,2015 Jaymin Patel, Heart Disease J48 56.76%
Prof.TejalUpadhyay, Dr. Prediction Logistic Model tree 55.77%
Samir Patel Using Machine algorithm
learning and
Data Mining
Technique

2 Mar,2018 V.V. Ramalingam, Heart disease Naïve Bayes 83.49%


Ayantan Dandapath, M prediction using SVM 84.81%
Karthik Raja machine K-NN 83.16%
learning Decision Tree 77.55%
techniques Random Forest 91.6%

3 Dec,2018 Amin Ul Haq , Jian Ping A Hybrid Logistic regression 87


Li , Muhammad Intelligent K-NN 85
Hammad Memon , Shah System Artificial Neural 86
Nazir , Ruinan Sun. Framework for network
the Prediction SVM 88
of Naïve Bayes 83
Heart Disease Decision tree 84
Using Machine Random Forest 83
Learning
Algorithms
4 May,2019 N. Satish Chandra Classification K-NN 81.9%
Reddy, Song Shue Nee, and Feature SVM 93.75%
Lim Zhi Min, Chew Xin Selection Random Forest 83.70%
Ying Approaches by Naïve Bayes 85.92%
Machine Neural Network 82.51%
Learning
Techniques:
Heart Disease
Prediction.
5 Jul,2019 S Anitha, N Sridevi. Heart disease K-NN 76.67%
prediction using Naïve Bayes 86.6%
datamining SVM 77.7%
techniques

6 Sep,2020 Devansh Shah, Samir Heart disease Naïve Bayes 83.49%


Patel, Santosh Kumar prediction using Decision Tree 71.43%
Bharti machine K-NN 83.16%
learning Random Forest 91.6%
techniques
7 Nov,2020 Malavika G, Rajathi N, Heart Disease Random Forest 91.8%
Vanitha V, Parameswari Prediction Naïve Bayes 88.52%
P Using Machine Decision Tree 78.68%
Learning SVM 88.52%
Algorithms KNN 86.88%
Logistic Regression 86.88%
8 Feb,2021 Pratiksha Shetgaonkar, Heart Disease Decision Tree 98.54%
Shailendra Aswale Prediction using Naive Bayes 85.01%
Data Mining Neural network 81.83%
Techniques

9 May,2021 Shruti Kothadia Classification Logistic Regression 85.05%


algorithms in Decision Tree
Python – Heart Random Forest 78.16%
Attack K-NN 86.20%
Prediction and 59.77%
Analysis

10 Aug,2021 Sibo Prasad Patro, Gouri Heart disease KNN 80%


Sankar Nayak, prediction by Naïve Bayes 86.7%
Neelamadhab Padhy using novel Neural Network 80%
optimization SSA-NN 86.7%
algorithm: A SVM 80%
supervised BO-SVM 93.3%
Learning
prospective.
Conclusion:

In the above it shows that machine learning techniques plays a major role in predicting diseases.

The most commonly used algorithms in the heart diseases prediction are as follows:

 SVM (Support Vector Machine)


 K-NN (K-Nearest Neighbor)
 Random Forest
 Decision Tree

In future by using optimization techniques and selection algorithm more experiment will be performed to increase
the accuracy of the prediction in heart disease.

You might also like