HybridDeepCardio: A CNN and Random Forest-Based Model for Heart Disease Classification

CVDs are currently the leading killers accounting for 32 percent of all fatalities in the world. This work employs sophisticated machine learning approaches to identify five main heart diseases: It can also affect the blood vessels, the heart muscles, and the heart valves through conditions such as CAD, heart valve disorders, heart failure, cardiomyopathy, and congenital malformation of the heart. The classification was performed using Random Forest methods while Convolutional Neural Networks were utilized for feature extraction from the dataset of 2265 photos of the medical nature annotated. Preprocessed images were split into the training, validation, and test data sets to guarantee complete evaluation of the model. With all the measures taken holistically showing remarkable performance in most of the disease categories, the hybrid model's general accuracy was 87%. Particularly for CAD, the model obtained the F1-values of the given data as 87. 34%. Such percentages show that the given treatment responded favorably in 96% of Heart Valve Disease patients and 90% of Peripheral Vascular Disease patients. This paper made use of information from 85% of Congenital Heart Disease. These presumptions show how effectively the model may improve the probability of disease diagnosis to support treatment, so the model is a vital instrument in the diagnosis of disease identification. The work underlines the need of using CNNs and Random Forest algorithms in medical diagnosis and opens the path for the developments of work in the field of education.