Heart Disease Prediction Using Machine Learning

Abstract— The heart plays an important part in living II. MACHINE LEARNING
organisms. opinion and prognostic of heart complaint must be Machine learning in healthcare involves the use of various
made easily, exhaustively and directly, because the fewest techniques like genetic algorithms, deep learning, and data
negligence can lead to serious complications or death,
mining to analyze large datasets and extract meaningful
numerous heart conditions are threat factors for death, and the
patterns. These patterns are then used to make predictions
number is gradationally adding . To break this problem,
prophetic styles that will ameliorate people's understanding of and decisions regarding the presence of diseases, including
the complaint are urgently demanded. Machine literacy is a heart disease. The advantage of using machine learning
branch of artificial intelligence( AI) notorious for furnishing algorithms in healthcare is that they can consider multiple
support in prognosticating any situation that requires training contributing risk factors simultaneously, enabling a more
from natural marvels. In this paper, we calculate the delicacy comprehensive and accurate prediction model.
of machine literacy algorithms for prognosticating heart
complaint, similar as k- nearest neighbor, decision tree, direct
retrogression, and support vector machines, through training
and evaluation using the UCI depository dataset( SVM).
Anaconda( jupytor) primer is the stylish tool to use Python
programming. It has colorful functions in the library and title
lines to make it more effective and accessible.

Keywords—Anaconda, Decision Tree

I. INTRODUCTION
Heart complaint is a major public health concern worldwide,
causing significant morbidity and mortality. Multitudinous
risk factors contribute to the development and progression of
heart complaint, making it challenging to accurately predict
its presence in individualities. In recent times, the use of
machine learning algorithms has shown promise in
developing prediction models for heart disease. These
algorithms have demonstrated the ability to analyze large
datasets and identify patterns and relationships that can aid in
predicting the presence of heart disease. Previous research
studies have utilized various machine learning techniques
such as support vector machines, logistic retrogression, naive
Bayes, and k-nearest neighbor algorithms to prognosticate
and classify heart complaint with varying degrees of .
accuracy.
Machine literacy algorithms, such as Support Vector
Machines, Naive Bayes, and K-nearest neighbors, have been
widely used in the prediction and classification of heart
In 2014, Dai et al. conducted a study using classification disease . These algorithms analyze various factors such as
models such as SVM, logistic regression, and naive Bayes, blood pressure, insulin level, cholesterol, pulse rate, and
achieving a prediction accuracy of 82% for heart disease. body mass index to identify potential risks and predict the
Similarly, in 2016, Kedar et al. used the KNN algorithm and presence of heart disease in individuals.
achieved an accuracy of 75% in predicting and classifying
heart disease . These studies indicate the potential of machine These algorithms are trained on large datasets that contain
learning algorithms in accurately predicting heart disease. clinical reports and laboratory test records of patients,
allowing them to learn patterns and make accurate
predictions based on these algorithms Selection for Heart
Machine learning algorithms have emerged as powerful tools Disease Prediction.
in healthcare for predicting and classifying various diseases,
including heart disease. These algorithms analyze clinical
reports, laboratory test records, and other patient data to
detect the presence of diseases such as diabetes, Alzheimer's,
and heart disease. One commonly utilized algorithm is the
Support Vector Machine, which can effectively detect the
presence of heart disease based on clinical and laboratory
data. Researchers have also explored the use of Probabilistic
Neural Network algorithms for heart disease prediction, as
demonstrated by Dessai et al. Due to the complexity of heart
disease and its risk factors, machine learning
A. Supervised Learning III. RELATED WORK
The heart is an important part of the human body and plays
Supervised learning is a fundamental concept in an important role in transporting blood and oxygen, which
machine learning where an algorithm learns to make are very important to the human body and therefore need to
predictions or decisions based on labeled training data. be protected from normal blockages. This is still an
In this approach, the algorithm is provided with a important issue. anatomy. people. health. Scientists are
dataset that includes input data and corresponding studying this problem. Therefore, many scientists are
output labels. The input data represents the information working on this. Heart disease should always be evaluated,
you want to make predictions on, while the output whether we are talking about diagnosis or heart disease
labels represent the desired or target outcomes. The prevention. Many fields, including artificial intelligence,
primary objective of supervised learning is for the machine learning, and data mining, are attracting attention.
algorithm to learn a mapping from inputs to outputs so This job.
that it can accurately predict the output for new, unseen
data. The performance of an algorithm depends on the variability
and bias of the data set. Predicting heart disease with
machine learning study by Himanshu et al. Naive Bayes
The process typically involves the following key steps: handles variance and bias better than KNN. knn suffers from
First, you collect and prepare a labeled dataset, fitting issues due to lack of bias and high variance, so knn
ensuring that you have a sufficient amount of data to does not work as expected. Using variables has many
train the model effectively. Then, you choose a advantages and disadvantages as you need less amount of
machine learning model, such as a decision tree, neural data to spend less time training and testing the algorithm,
network, or support vector machine, and train it using but there are also some disadvantages to using data on small
the labeled training data. During training, the model pieces of paper. The probability of asymptotic error depends
learns to identify patterns and relationships in the data, on the size of the data set, in which case unbiased algorithms
which will enable it to make accurate predictions. based on low variance work well. Decision trees are one of
the non-parametric machine learning algorithms, but as we
all know, there are many problems that can be solved by
Once the model is trained, it is deployed to make
removing constraints. Support Vector Machine is an
predictions on new data. The model takes an input and
algorithm with background in algebra and statistics that
generates an output or prediction based on the patterns
generates individual n-dimensional superprograms for data
it has learned from the training data. To evaluate the
classification.
model's performance, you compare its predictions to
the
actual, known outputs from a separate test dataset. The path to the soul is difficult and must be walked
Common evaluation metrics, such as accuracy, carefully. Otherwise it will lead to death. The severity of
precision, recall, and F1-score, are used to assess the cardiovascular disease is classified according to various
model's effectiveness in making prognostications. methods, including KNN, Decision Tree, Generalized
Supervised learning is widely applied in various Method, and Naive Bayes. Some researchers, including
domains, including Mohan, are working on collecting data to predict heart
image and speech recognition, natural language disease. Kaul et al review this and describe how to extract
processing, medical diagnosis, recommendation interesting patterns and information from large data sets.
systems, and more, making it a fundamental concept in Compare the accuracy of different machine learning and
the field of machine learning. data mining methods.

B. Unsupervised Learning
Unsupervised learning is a branch of machine learning IV. METHODOLOGY OF SYSTEM
where algorithms are utilized to analyze and find
patterns in datasets that lack explicit supervision or
The study began with data collected from the UCI repository
labeled data. Unlike supervised learning, which
and was completed by researchers and UCI representatives.
involves training a model to predict specific outcomes,
A. Data collection. The first step in system evaluation is to
unsupervised learning aims to uncover inherent
collect data and identify training and testing data. This
structures, relationships, or similarities within the data
project uses 73% of the training data and 37% of the dataset
itself. There are two primary categories of unsupervised
as testing data. rain. The attributes of the character selection
learning techniques: clustering and dimensionality
dataset are those used for body and mind, a person's heart
reduction. Clustering algorithms group data points into
rate, gender, age, and other information for prediction.
clusters based on their inherent similarities, without
prior knowledge of the categories or labels. Common
clustering methods include K-Means and hierarchical
clustering.
To obtain known results using machine learning algorithms,
prior information must be provided. For example, the random
forest algorithm does not support null data, so you need to
check the significance of the original data. In our project, we
need to test code for dummy values containing "0" and "1"
using: . Data Balancing Data balancing is very important to
achieve truth. Because data balance shows that these two goals
are balanced. Figure 3 represents the target group. Here, “0”
represents patients with heart disease and “1” represents
patients without heart disease.

Target Classes View

V. MACHINE LEARNING ALGORITHMS

A. Linear Regression