Download PDFOpen PDF in browserPersonalized Heart Monitoring and Reporting SystemEasyChair Preprint 52486 pages•Date: March 30, 2021AbstractTreatment cost for heart ailments is very high and there are people in India who can’t bear the cost of general treatment for their heart ailments. Indeed, even there are ones who live in remote places and can’t look for a decent specialist. All the customary advancements are presently being updated with the time of web and innovation. Innovation makes things look less demanding and advantageous to utilize. Along these lines, we intend to apply the Machine Learning algorithms and distinctive classifiers to anticipate the probability of heart diseases and altogether analyze them and give exercises and medications to counteract it. Heart disease is taken into consideration as one of the important reasons of death across the globe. Most of the cases include lack of money for routine check-ups, lack of transportation facilities, although most of the problems can be cured with the help of medications and exercises. With the motive to provide names of medications and appoint exercises for persons to stay healthy and eventually reduce the death toll caused by heart disease, we have analyzed and compared data mining techniques of Random Forest (RF), K-Nearest Neighbor (KNN), Gradient Boosting Machine (GBM), Generalized Linear Model (GLM). The performance of each of these algorithms was measured and compared with respect to factors like accuracy, confusion matrix, ROC curve, and AUC value. The best algorithm is then used to predict the probability of heart disease on the parameters provided by the user to generate a report with medications to prevent coronary illness and exercises to stay healthy and fit. Keyphrases: Generalized Linear Model, Gradient Boosting Machine, Heart Disease, K-Nearest Neighbor, Random Forest, machine learning
|