Evaluation of Machine Learning Models for Predicting Cardiovascular Disease Based on Framingham Heart Study Data

Abstract

The Framingham Heart Study Community is a research community that produces data related to Cardiovascular Disease (CVD). This research applies technology to predict CVD using machine learning based on data from the Framingham Heart Study. The eight machine learning algorithms are deployed in this study, they are decision tree, naïve bayes, k-nearest neighbors, support vector machine, random forest, logistic regression, neural network, and gradient boosting.This research uses several stages of research such as load dataset, preprocessing data, data modeling, evaluation of various data modelling, and input new data. The best performance was produced by the random forest model with an accuracy value of 0.84, a precision value of 0.84, a recall value of 0.85, an f1-score value of 0.79 and an AUC value of 0.72. The prediction generated by the proposed machine learning model is high risk or low risk of CVD.

Keywords

• machine learning
• cvd
• framingham heart study