Survival Analysis of Heart Failure Patients Using Advanced Machine Learning Techniques

According to World health organization, the death rate due to cardiovascular Disease (CVD) globally touches an estimate of 17.9 million every year. It became very significant for the doctors to accurately predict the heart diseases and failures based on various clinical reports across multiple patient records. Machine learning and Deep Learning techniques made a larger impact in making timely decisions by the doctors to suggest appropriate medication based on prediction models there by controlling the death rate due to diseases associated due to heart failure. Artificial Intelligence (AI) techniques have been more popular in the recent times for accurate clinical diagnosis in health care. This paper mainly concentrates in identifying the best machine learning technique for generating a predictive conclusion of the rate of survival in heart disease victims. A dataset comprising of 303 samples with 14 attributes of heart disease patients is considered for the analysis The contribution mentioned in the paper is to analyze and evaluate the best prediction model for identifying flow in the heart failure. The machine learning models considered for evaluation and comparison are Decision Tree, Random Forest, XG Boost and Ada Boost. It is found that Logistic Regression exhibits an accuracy of 0.885, XGBoost and Decision Tree exhibits equal performance with an accuracy 0.8524.