Integrated approach using deep neural network and CBR for detecting severity of coronary artery disease

Abstract

Despite major diagnostic progress and treatment progress, cardiovascular diseases (CVD) continue to be the world's leading cause of disease and mortality. Artificial intelligence methods provide the ability to drastically alter cardiology healthcare, by improving the reliability and optimizing the CVD prediction and response accuracy. Medical knowledge can also be improved by AI techniques like machine learning and depth learning due to the availability of healthcare data related relevant cardio clinical information. The focus of this research is to diagnose coronary artery disease among patients based on their clinical data using a deep neural network. The paper focuses on the dual approach where in the first phase diagnosis of coronary artery disease (CAD) is carried out using a deep neural network. The Deep learning-based model has achieved the highest prediction accuracy of 96.2% and a lowest error rate of 3.8 %. Further to handle the overfitting Gaussian noise is introduced into the model to improve the performance and in the second phase the severity of the disease is checked using case-based reasoning approach (CBR).

Ähnliche Arbeiten