Heart Failure Prediction Using Machine Learning Techniques

In this modern era people are very busy and working hard in order to satisfying their materialistic needs and not able to spend time for themselves which leads to physical stress and mental disorder. There are also reports that heart suffer because of global pandemic corona virus. Inflammation of the heart muscle can be caused by corona virus. Thus heart disease is very common now a day’s particularly in urban areas because of excess mental stress due to corona virus. As a result Heart disease has become one of the most important factors for death of men and women in the so called material world. It has emerged as the top killer that has affected both urban and rural population. CAD (Coronary artery disease) is one of the most common types of heart disease. In the medical field predicting the heart disease has become a very complicated and challenging task, requires patient previous health records and in some cases they even need Genetic information as well. So, in this contemporary life style there is an urgent need of a system which will predict accurately the possibility getting heart disease. Predicting a Heart Disease in early stage will save many people’s Life. There were many heart disease prediction systems available at present, the Authors have been researched well and proposed different Classification and prediction algorithms but each one has its own limitations. The main objective of this paper is to overcome the limitations and to design a robust system which works efficiently and will able to predict the possibility of heart failure accurately. This paper uses the data set from the UCI repository and having 13 important attributes. This work is implemented using many algorithms such as SVM, Naive Bayes, Logistic Regression, Decision Tree and KNN. It is found that SVM gave the best result with accuracy up to 85.2%. A comparative statement of all the algorithms also presented in the implementation part of the paper. This research also uses model validation technique to design a best suitable model fitting in the current scenario.