ECG - Driven Heart Disorder Profiler using Machine Learning Techniques

Right now, heart illness is the driving cause of passing around the world and its frequency is on the rise. To precisely analyze heart illness in its early stages some time recently a heart event occurs, it is troublesome to do so. There is a large amount of data on heart disease available in the healthcare sector such as clinics, hospitals, etc., but it is not easy to process this data to identify hidden patterns. Machine learning techniques can help to transform this medical data into valuable information. Machine learning can be used to design support system that can learn from their experiences. The aim of the research is to furnish the heart illness diagnosis in the most precise way. The proposed work uses machine learning and feature selection algorithms to compute results using the hybrid model. Methods like the Random Forest, Boruta, decision tree, logistic regression (LR), support vector machine (SVM), logistic regression with SVM, and hybrid models that combine logistic regression with SVM and Random Forest were used in this process. The assessment includes both individual and ensemble models and covers all heart disease datasets. Hybrid models are evaluated across multiple datasets using key metrics such as true positive rate, specificity, accuracy. The proposed framework's results are analyzed with following extensive experimentation. The study found that the proposed hybrid machine learning models outperform individual models and approaches in terms of accuracy, sensitivity, and specificity across all heart disease datasets.