Predicting Hypertension at an Early Stage Using Machine Learning Methods

Abstract

Hypertension poses a significant global health issue that frequently develops without symptoms, elevating the risk of severe cardiovascular illnesses. Timely prediction is crucial for prompt intervention and effective management. This research introduces a machine learning framework aimed at forecasting hypertension through the UCI Hypertension Dataset, which encompasses demographic, lifestyle, and clinical data from 1,500 individuals. The data preprocessing phase involved addressing missing values, employing one-hot encoding, normalization, and feature selection via Recursive Feature Elimination (RFE). Three supervised learning algorithms—Support Vector Machine (SVM), Random Forest, and XGBoost—were utilized. The objective of this framework is to enhance preventive healthcare measures and improve clinical decision-making by delivering accurate predictions regarding hypertension.

Ähnliche Arbeiten