Descriptive and Predictive Analysis on Heart Disease with Machine Learning and Deep Learning

Abstract

At present, one of the most fatal diseases in the world is heart disease. The mortality rate caused by heart disease is still relatively high, thus more intense effort in the prevention is needed, for instance by improving the achievement of a prediction model on heart disease. This research objective is to implement a prediction comparison of several machine learning and deep learning models, whether an individual is suffering from heart disease or not. In this research, the methodology used comprises of three machine learning models and three deep learning models to obtain the highest accuracy in predicting heart disease. Machine learning models applied in this research are Logistic Regression, Support Vector Machine (SVM), and Naïve Bayes, meanwhile models which are applied for Deep Learning method include Long Short Term Memory (LSTM), Convolutional Neural Network (CNN), and Recurrent Neural Network (RNN). The accuracy obtained from this research consists of Logistic Regression by 86%, SVM by 88%, and Naïve Bayes by 86%. Meanwhile, the accuracy of LSTM achieves 84%, RNN has 90%, and CNN takes 84%. The conclusion derived from this research is RNN model prevails with the highest accuracy by having 90% and becomes the best one to predict whether an individual is suffering from heart disease or not.