Multi-Model Supervised Machine Learning Techniques for Heart Stroke Prediction

This research presents a supervised machine learning approach for heart attack prediction using several models. Prompt care and preventative measures may be possible with early recognition of heart attacks, which pose a significant danger to health. This study proposes a new method that integrates the strengths of many machine learning models to improve the precision of the generated predictions. Blood pressure, cholesterol levels, age, gender, medical history, and other risk indicators are all part of the study's dataset. The dataset is preprocessed for learning the missing values and standardizing the features. Three separate ML algorithms-logistic regression, random forest, and support vector classifier (SVC)-are then trained by using the preprocessed data. The prediction ability is enhanced by stacking using the differences between each model. Then, performance metrics, including F1-score, recall, accuracy, and precision are used to evaluate the stacked model. The results of the experiments demonstrate that the multi-model approach outperforms the individual models, leading to more reliable and precise predictions. The proposed technique shows the potential to predict heart attacks and assist doctors in identifying people at high risk of experiencing a heart stroke.