Prediction of Heart Failure Using Random Forest and XG Boost

Abstract

Heart Failure (HF) is a prevalent illness that can lead to hazardous situations. Each year, approximately 17.9 million patients globally die of this disease. It is challenging for heart specialists and surgeons to predict heart failure accurately and on time. Fortunately, there are classification and prediction models available that can assist the medical field in efficiently using medical data. The objective of this study is to enhance the accuracy of HF prediction by using a Kaggle dataset composed of five sets of data over 11 attributes. Multiple machine learning approaches were utilized to understand the data and forecast the likelihood of HF in a medical database. The results and comparisons show a definite increase in the accuracy score of predicting heart failure. Integrating this model into medical systems would be beneficial for predicting heart disease using data collected from patients.