A Prediction Model for Heart

Disease Patients

Guided by : Done by:

Dr.Manikandan G
R. Hindhuja(120015078)
G. Chakradhara Rao(120003058)
A. Tejasri(120014004)
18-11-2019 1
Table of contents
 Base paper
 Data mining- Introduction
 Heart diseases-Introduction
 Abstract
 Architecture
 Algorithms
 Data set description
 Attribute description
 GUI

18-11-2019 2
Base Paper:
 Anam Mustaqeema, Syed Muhammad
Anwar, Abdul Rashid Khanb,Muhammad
Majid, A statistical analysis based
recommender model for heart disease
patients , International Journal of Medical
Informatics 108(2017) 134-145.

18-11-2019 3
Data Mining Introduction:
Data mining is a process of extracting interesting
patterns from large sets of data that can be used for
predictions.
In today’s world data is everything , it plays a vital
role in business analysis but it has it’s own
prominence in medical field.
To extract features from huge sets of data, as per
the given attributes we use data mining techniques.

18-11-2019 4
Heart disease Introduction:

• Cardiovascular diseases are one of the major

causes of death in the last two decades.
• In all age groups we can identify this problem
in alarming rate
• In recent years, health monitoring systems
have evolved to be used in a more
challenging manner
• The main aim here is to predict the
presence of disease.

18-11-2019 5
Abstract:
• This project provides a prediction model that
enhances disease prediction to cardiac patients.
• The prediction model identifies the disease of a
patient and classifies it into two classes i.e.,
Heart Disease , No Heart Disease

18-11-2019 6
Architecture:

18-11-2019 7
18-11-2019 8
Algorithm (HD_PM) Model
Input: Patient's clinical dataset D
Output: Disease prediction P=0,1

1. Let D be divided into Training (Dtrain) and Testing sets (Dtest)

2. foreach Patient's data from Dtrain, do
3. if Dtrain has missing values then,
4. Apply pre-processing steps.
5. End
6. Let F. S represent selected feature set from Dtrain,
F. S = (f 1, f 2, f 3, f 4, f 5, f 6, f 7)
7. End
8.TrainClassifier C = (F. S, Datalabel)
9. Disease Prediction P = (C, Dtest)
10. Return Disease prediction P

18-11-2019 9
Dataset description
 collected from UCI machine learning repository
 14 attributes
 The num field refers to the presence of heart disease in
the heart patient integer valued from 0 (no presence)
to 1

Dataset

18-11-2019 10
Attribute description
1. age - age in years
2. sex-sex (1= male; 0 = female)
3. cp-chest pain type
4. trestbps-resting blood pressure (in mm Hg on admission to
the hospital)
5.chol-serum cholestoral in mg/dl
6. fbs-fasting blood sugar > 120 mg/dl(T,F)
7. restecg-resting electrocardiographic results
8.thalach- maximum heart rate achieved
9. exang-exercise induced angina (1 = yes; 0 = no)
10. oldpeak-ST depression induced by exercise relative to rest
11. slope-the slope of the peak exercise ST segment
12. ca-number of major vessels (0-3) colored by flourosopy
13. thal-3 = normal; 6 = fixed defect; 7 = reversable defect
14. num (the predicted attribute)

18-11-2019 11
The Classifiers used are:
SVM
 provides a hyper plane to classify the labelled data.
 a binary classifier
 For M number of classes, this classification gives (M -1)(M)/
2 number of binary classifiers

Random Forest
 a large number of individual decision trees that operate as an
ensemble
 It is similar to bagging, but when the classification tree is
growing, random forest considers only the binary split points

18-11-2019 12
Contd:
 MLP
 It has more than one perceptron and has two phases.
 First is forward pass, where the classified outputs for the
given input data are evaluated by

 Then back-propagation is used to make those weigh and bias

adjustments relative to the error and error itself can be
measured by using RMSE

18-11-2019 13
Contd:

Gradient Boosting
 a loss function to be optimized
 a weak learner to make predictions
 an additive model to add weak learners to minimize the
loss function
Decision Trees
 most widely used classifier for supervised learning
 used for both classification and regression
 generate understandable rules inferred from the data
features

18-11-2019 14
Contd:
Logistic Regression
 classification algorithm used to assign observations to a
discrete set of classes
 Unlike linear regression which outputs continuous number
values, logistic regression transforms its output using the
logistic sigmoid function
 returns a probability value which can then be mapped to two
or more discrete classes

18-11-2019 15
contd:
RANDOM FOREST

18-11-2019 16
Code:
SVM

18-11-2019 17
Decision Tree

18-11-2019 18
Gradient Boosting

18-11-2019 19
LOGISTIC REGRESSION

18-11-2019 20
contd:
MLP

18-11-2019 21
GUI

18-11-2019 22
GUI

18-11-2019 23
GUI

18-11-2019 24
GUI

18-11-2019 25
GUI

18-11-2019 26
GUI

18-11-2019 27
GUI

18-11-2019 28
Thank You

18-11-2019 29