Prediction of Hypertension Disease Using Machine Learning Techniques with Random Committee Ensemble Learning

Hypertension is still among the most common long-term illnesses in the world, killing about 8.5 million people each year and resulting in 13% of worldwide deaths. The number of cardiovascular deaths is on the rise, especially throughout South Asia, where 88% of such fatalities take place in low- and middle-income countries, attributed to greater availability of unhealthy foods, sedentary lifestyles, as well as rural-urban migration trends. Conventional diagnostic approaches that depend on intermittent clinical assessments as well as patient-reported symptoms frequently do not facilitate prompt recognition of at-risk individuals, leading to postponed treatment and the advancement to serious cardiovascular complications. The present investigation presents a comprehensive machine learning framework for the early prediction of hypertension, utilizing Random Forest (RF) and Random Committee (RC) classifiers assessed on a primary dataset especially gathered for this research. The suggested Random Committee ensemble method works by making several base classifiers, usually decision trees, each one trained with the same dataset but starting using various random seeds to add variety to the algorithms. Using five standard metrics to evaluate performance shows that both the Random Committee and Random Forest models have been more accurate (RF:$\mathbf{9 2. 3 4 \%, ~} \mathbf{R C: ~ 9 3. 6 8 \%}$) than traditional statistical methods. This Random Committee model shows the best balance between reducing variance and making predictions stable by means of majority voting aggregation.