A Mixed-Methods of ML Strategy for Cardiovascular Disease Prediction Using Classification Data

Heart attack and stroke poses a threat to public health due to its high death rate and widespread occurrence. The era of early heart illness prediction based on a handful of straightforward physical indicators gleaned from standard physical examinations is passed. In order to establish precise forecasts and make genuine attempts towards future diagnosis, signs of cardiac sickness are very responsive and clinically important. It takes a lot of effort and time to manually examine and predict a lot of data. This study aims to develop a new model that can quickly and accurately predict heart disease using a collection of biological signs. Our cardiovascular disease forecasting system evaluates the predictive algorithms’ ability to classify on mixed datasets using the train-test split technique. Previous studies are contrasted with the proposed method’s training results. Various approaches to comparative analysis and their successful outcomes are explored in this work. The results of this study show that ML algorithms perform better than their human counterparts. Accuracy, recollection, F1-value, misperception matrices, and ROC-AUC curvatures are key performance metrics among the comparative methodologies used. The heart illness dataset yields the best results when using the suggested mixed ML approach, with F1-score, accuracy, precision, and recall all at $\mathbf{1 0 0 \%}$.