Predicting The Occurrence Of Cardiovascular Disease Using The Novel Ensemble And Blend-Based Networks, EnsCVDD-Net and BlCVDD-Net

Abstract

Cardiovascular disease (CVD) is the most motive of demise globally, requiring well timed and unique analysis for effective medical intervention. this technique employs modern computational methodologies, using deep learning (DL) and machine learning (ML) strategies to enhance the accuracy and resilience of cardiovascular ailment (CVD) prediction systems. In contrast to standard machine learning strategies that rely notably on manual feature engineering, deep learning models possess the potential to independently extract complicated features from unprocessed records, rendering them noticeably appropriate for complex scientific datasets. The Adaptive synthetic Sampling (ADASYN) approach is applied to tackle troubles such as class imbalance, enhancing minority class representation and facilitating balanced learning. A dataset referring to heart disease is applied to educate and check the suggested class framework. Two novel ensemble-based architectures, namely EnsCVDD-Net and BlCVDD-Net, are introduced—one leveraging traditional ensemble strategies like bagging and boosting, and the other employing blend-based meta-learning techniques to combine multiple base classifiers effectively. The methodology incorporates major ML algorithms such as Random Forest (RF), XGBOOST, LightGBM, and Deep Learning models, including CNN and LSTM, to provide superior function representation. An efficient voting classifier for aggregate predictions from several models is also implemented, which improves decision-making accuracy. Metrics of evaluation, such as accuracy, download, and F1-score, demonstrate the advantages of file access over individual models. The system achieves 91.7% classification accuracy, 92.0% accuracy, 91.7% dismissal, and 91.8% F1 score, suggesting balanced performance across all rating metrics. The results highlight the adaptability and reliability of EnsCVDD-Net and BlCVDD-Net across varied clinical scenarios, reinforcing their potential in real-time medical diagnostic systems. This strategy underscores the promise of intelligent systems in delivering high-accuracy, automated predictions for early-stage cardiovascular disease detection.