Survey of Heart Disease Prediction Using Various Machine Learning Technique

Abstract

Many prediction techniques targeted at early identification and intervention have been developed since heart disease continues to be a major cause of death worldwide. The predominant approaches used in cardiac disease prediction are examined in this review of the literature, with an emphasis on statistical models, hybrid approaches, and machine learning (ML) techniques. Because they provide comprehensible findings and are often easy to execute, traditional statistical approaches like logistic regression and decision trees have served as the foundation for predictive modelling. Through the use of sophisticated algorithms like Random Forest, Support Vector Machines (SVM), and Neural Networks (NN), machine learning has, nevertheless, dramatically improved forecast accuracy. Recent research highlights deep learning methods that enhance early diagnosis by recognizing complex patterns in large datasets. Furthermore, the integration of clinical data, such as ECG signals, patient history, and genetic markers, with ML algorithms is increasingly recognized for its potential to elevate prediction accuracy. Challenges persist, particularly regarding data quality, model interpretability, and the clinical application of advanced models. The review concludes by highlighting the importance of personalized models and real-time prediction systems in future research, aiming to bridge the gap between algorithmic development and clinical utility.