CardioML: Performance Evaluation of Intelligent Prediction of Heart Attacks using ML and DL Algorithms

Predicting heart disease during its preliminary stage remains vital for enhancing three key aspects: patient recovery rates, healthcare costs and individualized treatment approaches. The research evaluates multiple heart disease datasets through the use of Logistic Regression, Random Forest, Support Vector Classification (SVC), Gradient Boosting, Decision Tree, Naive Bayes and Deep Learning ML models. The presented datasets represent different parts of left ventricular hypertrophy heart divisions. The research evaluates the predictive abilities of those algorithms alongside their sensitivity, specificity and predictive accuracy as well as their robustness for dealing with medical research datasets that display extensive, scattered, and unbalanced patterns. The ensemble methods, Random Forest and Neural Network (Deep learning) in machine learning, provide extremely accurate results based on the research findings that exceed traditional statistical models in different scenarios. The study focuses on machine learning difficulties in healthcare that involve protecting medical data as well as training diverse large datasets.