Cardiovascular Disease Prediction Model using Machine Learning Algorithms

Abstract

A general term for conditions affecting the heart or blood vessels is called as Cardiovascular disease (CVD). It is commonly associated with an increased risk of blood clots and build-up of fatty deposits inside the arteries (atherosclerosis). Sometimes, it can also be associated with damage to arteries in organs such as the brain, kidneys, heart and eyes. CVD is the reason for the highest number of deaths globally and the major cause of death annually. Most cardiovascular diseases can often be prevented by leading a healthy lifestyle and addressing behavioural risk factors such as unhealthy diet and obesity, tobacco use, harmful use of alcohol and physical inactivity using population-wide strategies. Machine Learning can play an important role in predicting cardiovascular disease and such information, if predicted well in advance can provide significant insights to doctors who can then adapt their treatment and diagnosis for each patient accordingly. In the proposed research method, firstly the attributes are selected from the dataset, then data pre-processing takes place which uses techniques such as removal of noisy data, removal of missing data, filling default values if applicable, classification of attributes for prediction and decision making at different levels. Classification, accuracy, sensitivity and specificity analysis is done to obtain the performance of the diagnosis model. A prediction model which predicts whether a person has a heart disease or not and hence provide diagnosis or discussion on the results is proposed. This is accomplished by applying rules to the individual results of classification algorithms such as Gradient Boosting Classifier, Random Forest Classifier, Support Vector Machine, Extremely Randomized Trees Classifier (Extra Trees Classifier), Logistic Regression and Multi-Layer Perceptron (MLP) Classifier obtained on the dataset.

Ähnliche Arbeiten