Hypertension Classification Using HistGradientBoostingClassifier, HealthD, And Model Optimization

High blood pressure ranks among the world's most common heart-related conditions, carrying serious dangers like strokes and heart attacks. Even with progress in medical testing, spotting it early is tough because of the intricate mix of daily habits and inherited traits. This study seeks to solve the issue of precise hypertension forecasting using machine learning methods tailored for varied health information. Driven by the rising demand for evidence-based health prevention, the research employs the HistGradientBoostingClassifier on a collection of 1,985 patient profiles with eleven lifestyle and bodily indicators, such as age, body mass index, sleep hours, sodium consumption, and tension levels. The key innovation here is the histogram-based boosting approach, which adeptly manages diverse attributes and curbs excessive fitting via timely halting and adjustment techniques. Assessment findings show the model reaches 97% accuracy, maintaining even performance in precision, recall, and F1-score for both hypertensive and non-hypertensive groups. These findings underscore the model's reliability and suitability for inclusion in prompt alert tools for hypertension danger assessment. Upcoming efforts will investigate model clarity through SHAP analysis and pit boosting classifiers against neural network methods to boost understanding and adaptability in practical medical settings.