Main Article Content
Here it is a necessity for physicians to foresee heart problems before they see them in their cases. The factors that contribute to the odds of heart diseases are smoking, absence of physical work regime, hypertension, elevated cholesterol, improper diet, alcohol consumption, obesity, and sugar level variation in the body, etc. Heart related disease clinically is known as cardiovascular disease. Many factors and reasons influence the heart disorder, and it is the one of fundamental cause of death worldwide. Since, cardiovascular related ailments and diseases can be diagnosed through various clinical test results, these factors that leads to the test results give an opportunity to perform analysis on them and enable to predict the cardio disease at an early stage by identifying the contributing risk factors and treating them to prevent fatality. There are techniques like Data mining approaches and machine learning (ML) algorithms like Naïve bayes, Random Forest, Decision tree, Support vector machine, K-nearest neighbor, Convolutional Neural network, these will facilitate predicting the current condition of heart health. ML is the process of data evaluation from different perspectives and combines useful information. Machine
learning technique is used for attempting to predict the future scenario by training the machine.
This research work emphasis more on study and analysis of the Convolution Neural Network approach, Random Forest techniques, Naïve bayes, decision tree and support vector technique’s applicability for predicting heart health by anticipating the heart disease in patients. The reason behind choosing these two techniques is that majority of the research work carried out earlier contains these two techniques in majority and tried to analyze the performance difference between them through executing these algorithms on the selected dataset. To solve the complexities in identifying the heart disease and a decision support system, is based on the machine learning algorithms such as, Support vector Machine, k-nearest neighbor (k-nn), Convolutional neural network, random forest, naïve bayes, decision tree, because of this reason we have chosen these algorithms for comparison with proposed model. With help of comparison came know that convolution neural network is best classifier for existing heart dataset. Performance of these algorithms like support vector, k-nearest neighbor (k-nn), support vector, random forest, naïve bayes, decision tree calculated through factor like f-measure, accuracy, precision, recall, execution time.