Predicting Chronic Heart Disease

Predicting Chronic Heart Disease
By
Author

Table of Contents
1.0. Introduction......................................................................................................... 1
1.1. Data set........................................................................................................... 1
1.2. Exploratory of the data....................................................................................... 2
2.0. Machine learning implementation............................................................................3
2.1. Collaborative filtering......................................................................................... 3
2.2. Logistic regression............................................................................................ 4
2.3. K-means.......................................................................................................... 6
3.0. Conclusion and summary...................................................................................... 7
References................................................................................................................... 8
Page i of xiv

1.0. Introduction
According to National Cancer Institute, (2016) chronic heart disease refers to a
heart condition causing it to have problem pumping blood throughout the
body. Some of the symptoms may include having short breath, fatigue,
swelling of abdomen. This disease may be caused by heart attacks, coronary
artery problems, and high blood pressure among others. The World Health
Organization has reported that approximately 12 million across the world has
lost their lives because of heart disease. WHO further pointed out that half of
the lives lost due to heart disease came from developed countries.
Discovering the heart disease during its early stages helps the patients to be
able to reform their lifestyle early on in order to prevent the risks that come
with it and reduce complications that come with the disease. There are a
number of risks that are connected to this disease. This research paper will
use some of the risks that leads to the occurrence of a heart disease such as
the blood pressure (such as systolic and diastolic blood pressure), cholesterol
in the body. These risks will be analyzed and computed in order to predict
whether or not there will be occurrence of a heart in about ten years. This
can help someone to know that their current lifestyle needs to be change or
maintained.
1.1. Data set
The used for predicting chronic heart disease is public and available on
https://www.kaggle.com/amanajmera1/framingham-heart-study-dataset
website. The data is from 2 years ago however the study is still going on for
the town people of framinghan in Massachusetts. The study is on
cardiovascular in order to prove whether or not it is possible to predict the
factors and the kind of lifestyle that may cause chronic heart failure and
whether or not studying these risks can really help to predict if a person will
have a cardiovascular disease in 10 years’ time. The file containing this
dataset is called Framingham.csv. The dataset has 4240 rows and 16
columns. The 16 attributes are shown below
Student name Page 1 of 14

The highlighted ‘ten year chronic heart disease’ (TenYearCHD) which has
binary data of 1 or 0 showing either true or false is the predicted attribute in
the dataset.
1.2. Exploratory of the data
The dataset has an education attribute which does not help in the prediction
and so dropping the column was necessary. Furthermore the dataset has
rows with null approximately 488 and since the number is very small
compared to the rest of over 40000, dropping the rows was insignificant.
Some of the graphs during exploration are shown below.
Figure 1 short code to get distribution of CHD patients
Source
Figure 1 above shows the number of people who have chronic heart disease
and thus that do not. 1 represents positive CHD and 0 represents negative
CHD. Figure 2 shows a bar graph to represent this phenomenon.
Student name Page 2 of 14