Least Absolute Prone Factor Based Cardio Vascular Disease Prediction Using DenseNet Multi Perceptron Neural Network for Early Risk Diagnosis

Heart disease poses a serious threat, and the occurrence of a heart attack can lead to premature and fatal situations. Recent investigations in data mining techniques have delivered valuable insights into analyzing cardiac data through sensing, monitoring, and learning data for early diagnosis. Artificial Intelligence (AI) is significant in acquiring accurate diagnoses through Machine Learning (ML) feature comments. However, current methods often must be remembered to accurately predict consequences due to imprecise features and inappropriate support weights when selecting components during training and validation. We propose a minimally exhaustive prediction-based approach for detecting early cardiovascular disease (CVD) risks using a dense multi-perceptron neural network to address this issue. The Cardiac Risk Impact Rate (CRIR) is estimated based on mean weight margins through decision tree attention to choose the cluster risk thresholds. The clusters are further evaluated using the Least Impact Disease Prone Rate (LIDPR) to observe feature margins accepted by Cardiac Frequency Limits (CFL). Based on the estimation rate, spider ant colony optimization selects the features by getting the active class margins. Then, the established cluster feature limits are trained into a Dense Net-Multi Perceptron Neural Network (DN-MPNN) to classify the risk levels. The early diagnosis was based on the risk levels to recommend the patients be safe from cardiac arrest. In addition, generating more precision than other methods can improve the accuracy of predictions. Furthermore, it retrieves relevant features based on deep feature data learning models to achieve high-impact projections of heart failure rates.