Heart Disease Prediction Model based on Prophet

Abstract

Heart disease is one of the major causes of death for people of all races, genders, and nationalities. In the United States, for instance, heart disease causes more than 600,000 deaths every year and is the largest leading cause of death in 2020. A reliable heart diseases mortality prediction model could acknowledge the patients’ medical professionals that the heart disease risk level of the specific group. This approach is significant in preventing further increases in heart disease mortality rates worldwide. Nowadays, multiple Machine Learning (ML) models, including hybrid models produced impressive predictions and realized that newly developed ML models might provide new perspectives on heart disease predictions. In this paper, we introduced the Facebook Prophet model (FB Prophet model), a time series prediction tool that could present seasonality in its result, since studies point out that heart disease mortality rate also shows seasonality. We produced an accuracy of approximately 94 % in predicting weekly heart disease mortality numbers in specific states. Furthermore, we explored the effects that external factors, ambient temperature, have on heart disease, and utilize this relationship in improving model accuracy.