Design and Development of Decision Support System for Reliable Prediction of Heart Disease using Machine Learning Techniques an Empirical Study and Analysis

Abstract

newline Heart disease (HD), also known as cardiovascular disease (CVD), refers to a group of disorders that impact the heart and blood arteries. It stands as a prominent global cause of mortality, with risk elements encompassing obesity, diabetes, smoking, elevated blood pressure, high cholesterol, and a sedentary lifestyle. HD can manifest in various forms, including coronary artery disease, heart failure, arrhythmias, and valve disorders. Given that HD typically progresses gradually, it is crucial to address risk factors and embrace a heart-healthy lifestyle to manage the condition. Timely identification and prevention are pivotal. HD prediction has become increasingly important in healthcare, with the help of advanced technologies and predictive models. The medical field has seen an increased emphasis on predicting heart disease due to advanced technologies and predictive models. Through the application of predictive technologies, healthcare professionals can assist individuals in reducing their risk of heart disease by offering proactive interventions and guidance, which may include medications and lifestyle modifications. Early detection and prevention remains the key aspect in the fight against heart disease, and predictive models play an important role in this effort by identifying high-risk individuals and enabling timely interventions to improve heart health. The primary aim of this research work is to design and evaluate a specific model for HD predictions with various classification algorithms. . This thesis represents notable advancement in the field of diagnosing and forecasting cardiac conditions. It introduces an innovative approach for early heart disease identification and medication administration by leveraging ensemble and deep learning techniques. Additionally, the work presents an efficient approach for early HD prediction by employing data pre-processing techniques, including oversampling and outlier removal. The methods are implemented and validated using an ensemble approach withou