PREDICTION OF CARDIOVASCULAR DISEASE USING GRADIENT BOOSTING

Information Technology

The modern civilization has extreme edge of advancements in the medical industry which acts like a backbone of evolution and technology. This decade of regime had a count of estimated 17.5 million deaths by 2012 with cardiovascular diseases. CVD is considered to be one of the main reasons of death all around the world. The prediction of CVD at an early stage may lower these rates. Heart monitoring is one of the most important aspect of healthcare monitoring. For detecting risk indicators, machine learning algorithms are a potential way. To acquire accurate cardiac disease prediction, this work introduces a gradient boosting (GB) approach. The aim of this project is to use the Machine Learning (ML) technology on medical data to predict the chance of CVD. The medical data available these days is often a valuable resource of quality information. This project aims to use supervised machine learning algorithms like Gradient Boosting, Logistic Regression, Decision Tree to obtain a CVD classifier. However, the challenge usually lies in the quantum of the data since there aren’t many datasets available regarding CVD, the correlations and the accuracy of the model when it comes to prediction using conventional techniques. This project proposes the use multiple datasets available in UCI repository namely Cleveland, Hungary, Statlog, Long Beach, Switzerland in deriving classifiers using above mentioned algorithms. The classifiers obtained using the datasets are then used to build a web application for users for them to know whether they are vulnerable to CVD or not based upon the details provided by them as input i.e., features that are used in classification of CVD. Depending on the prediction of the model the output is provided to the user, according to which the user should undergo lifestyle changes. The final goal is identification of the possibility of heart disease that can reduce the risk of mortality from cardiovascular disease (CVD).

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.