ABSTRACT
Nowadays, collected data history can be utilized to foresee potential patterns and assist businesses in making competitive decisions that will improve their success and benefits. Many analysts use data from the healthcare sector to discover and forecast ailments in order to help patients and doctors in a variety of ways. Cardiovascular diseases are one of the most serious ailments for humans, resulting in millions of deaths worldwide. According to the research performed by the World Health Organization, cardiac illnesses are expected to have caused 17.9 million deaths in 2019, accounting for 32% of all fatalities worldwide and with an annual mortality rate of more than 17.7 million. Of these fatalities, 85% were due to heart attack and stroke. Many researchers are working to create an effective strategy for the timely identification of heart disorders, as existing heart disease diagnosis methods are inadequate in early detection for a variety of reasons, including accuracy and computational time. Diagnosing and managing a cardiac disease is exceptionally challenging while modern technology and healthcare specialists are unavailable. While utilizing present algorithms, the results of this approach to diagnosis need to be revised in recognizing heart disease patients. The goal of this study is to ascertain how to diagnose and estimate the presence of cardiac disease. This chapter proposes a novel ensemble-classification-based approach in order to predict the presence or absence of heart illness using a few health-related factors while also enhancing the model’s accuracy. Certain of the patient’s health characteristics are selected on the basis of their importance in forecasting cardiac disease in this study. Several advanced learning techniques are utilized to improve the performance of the model. Comparative analysis of a few supervised classifiers (logistic regression, naïve Bayes classifier, Support Vector Machine algorithm, k-Nearest Neighbor algorithm and Decision Tree classifier) with the proposed model have been performed to illustrate how effective the ensemble learning technique is in improving the accuracy measure rather compared to the other algorithms. The suggested approach is trained on the Cleveland dataset from Kaggle and is tested using the meta-learner itself, resulting in the model being the most promising and evolutionary in the healthcare industry. While executing on the medical dataset, the ensemble classification’s performance was improved, resulting in 95.51% accuracy and 1.73 s of computational time.
