Hybrid Knowledgeable System for Life-style Recommendations in Hypertensive Sufferers

Abstract

This research presents a hybrid medical decision-support system for producing customized life-style suggestions in hypertensive sufferers by integrating unsupervised machine studying with rule-based knowledgeable reasoning. Using a real-world dataset of 615 medical data from routine healthcare providers, a preprocessing pipeline combining information imputation, normalization, and dimensionality discount was utilized previous to affected person stratification. Principal Component Analysis (PCA) preserved the dominant latent construction of the information, adopted by Ok-Means clustering, which recognized three clinically interpretable affected person profiles with a Silhouette coefficient of 0.5608. The clustered profiles had been built-in right into a rule-based inference engine structured throughout six life-style intervention domains: bodily exercise, stress administration, diet, sleep patterns, therapeutic adherence, and normal well being behaviors. Recommendations had been generated utilizing a dual-weighting technique prioritizing particular person affected person attributes whereas incorporating cluster-level contextual info. The system was evaluated via blind knowledgeable validation involving cardiologists and medical nutritionists. Agreement between automated suggestions and medical consensus reached 78.3%, with a Cohen’s Kappa coefficient of 0.742, indicating substantial concordance. No statistically vital variations had been noticed between system outputs and knowledgeable judgments (χ² = 8.347, p = 0.908). The outcomes show that combining unsupervised affected person stratification with express medical reasoning allows interpretable and scalable determination help for non-pharmacological hypertension administration, significantly in healthcare environments with restricted labeled information