International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 05 | May 2023
p-ISSN: 2395-0072
www.irjet.net
Prediction of Heart Disease Using Machine Learning and Deep Learning Techniques. Spurthi S R1, Shree Subeeksha G2, Keerthi M S3, Kavya P4, Dr. Gururaj Murtugudde5 1, 2, 3, 4Student, Dept. of Computer Science and Engineering, REVA University, Bangalore, India 5Professor, Dept. of Computer Science and Engineering, REVA University, Bangalore, India
---------------------------------------------------------------------***--------------------------------------------------------------------information to identify patients with heart disease, forecast Abstract - The primary cause of death has historically been
future heart illness, and identify it early. Different types of heart problems can be diagnosed, detected, and predicted with the aid of machine learning methods like Adaboost, ExtraTree algorithms, and deep learning techniques like CNN and Multi-layer Perceptron (MLP). The basic risk factors for heart disease are universal across the range of heart illnesses, enabling patients to get appropriate care and avert negative outcomes. To identify hidden patterns and analyse data to identify heart sickness at an early stage and prevent consequences, machine learning is essential.
heart related disease worldwide over past few decades, thus it is crucial and worrisome to anticipate any such disorders. Heart-related disease diagnosis and prognosis is a difficult task that calls for greater accuracy, correctness, and perfection since a small error can result in weariness or even death, which has a significant global impact. Due to a multitude of risk factors, such as smoking, diabetes, high cholesterol, and similar conditions, it can be challenging to diagnose heart disease. As a result of these circumstances, it is urgent to develop precise, practical, and trustworthy methods for making an early diagnosis, as doing so will benefit people everywhere by enabling them to receive the necessary therapy before the condition worsens. The data from the dataset is obtained using contemporary methods like data mining and machine learning techniques, and the fetched data is then utilised to forecast cardiac disease. With the help of deep learning techniques like CNN and MLP as well as machine learning methods like ADABOOST and EXTRATREES, this work attempts to predict the likelihood of getting cardiac illnesses.
2. LITERATURE REVIEW In this study, Kuldeep Vayadande et al. [1] used the 303-row and 14-attribute UCI heart dataset and implemented ml algorithms such as logistic regression, XGBoost, and random forest, which had good accuracy ratings of 88.52% in comparison to all other models. The accuracy of deep learning algorithms like MLP and ANN is 86.89% and 85.25 percent, respectively.
Key Words: Machine Learning, Adaboost, ExtraTree,
In a study to predict cardiovascular illness, Shafique R et al. used a variety of classification methods, such as Extra Tree Classifier, Logistic Regression, SVM, and NB The Cleveland heart dataset, which has two classes and 13 attributes, was used. The research discovered that, out of all the machine learning methods examined, Extra Tree Classifier achieved the highest accuracy rate of 90%.
Deep Learning, CNN, MLP, Heart disease Prediction.
1. INTRODUCTION The most important crucial organs in the human body, the heart plays a crucial part in the blood's circulation throughout the body. Heart disease can be caused by a number of things, such as unhealthy lifestyle choices, drinking, smoking, and job-related stress. This syndrome may result in abnormal heart blood flow and led to severe conditions such as strokes, coronary heart disease, and heart attacks. If the heart is faulty in any manner, cardiac disorders such congenital heart disease, heart failure, and arrhythmia may also manifest. Heart disease, is major cause for illness and loss of lives globally, killing 12 million people each year according to the World Health Organisation. Predicting cardiovascular disease is so essential, and many researchers have investigated the most important risk factors to precisely determine overall risk. The prevention of heartrelated illnesses depends on heart illness being detected early.
A. Lakshmanarao et al. predicted the risk of developing a coronary heart disease for a period of ten years in patients using Framingham Heart Study dataset. To solve the issue of an unbalanced dataset, three distinct sampling techniques were used using the dataset, which had 15 features. Using random oversampling, the study discovered that SVM was the most accurate machine learning model, while using adaptive synthetic sampling and synthetic minority sampling, ExtraTree and Random Forest were determined to be the most accurate models. This study reveals how ml techniques forecast a person's risk of getting heart disease, which may help with the early detection and treatment of this common condition. Shadab Hussain et al. recommended a 1D convolutional neural network (CNN) architecture to detect cardiac disease. The Cleveland dataset was used to train and test the model,
The enormous data created by medical industry is used by machine learning algorithms to make predictions and judgements. One such application is the analysis of patient
© 2023, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 91