Cardiac disease detection using cuckoo search enabled deep belief network
P. Nandakumar, Subhashini Narayan
Abstract
Cardiac disease is the most infected disease in the world nowadays for all ages of people. An emergency need arises to predict cardiac disease accurately in a short time. In this article, hamming distance feature selection method is proposed for the data preprocessing and data cleaning process in different cardiac disease datasets. Deep learning model such as deep belief networks is used with cuckoo search bio-inspired algorithm for finding the accurate prediction of cardiac disease. The results demonstrate that deep belief networks with the cuckoo search algorithm have achieved good performance with an accuracy of 89.2% from Cleveland, 89.5% from South Africa, and 89.7% from Z-Alizadeh Sani, 90.2% from Framingham, and 91.2% from Statlog cardiac disease datasets.