Healthcare AI

Inadequate access to timely healthcare guidance, especially in rural and underserved areas, contributes to delayed diagnosis and preventable health complications. This paper presents Healthcare AI, a comprehensive machine learning- based framework that integrates disease prediction, medicine recommendation, personalized diet planning, and daily routine generation into a unified web-based application. The system accepts multi- factor inputs including patient symptoms, age, BMI, body temperature, and comorbidities and leverages three supervised classification algorithms — Logistic Regression, Decision Tree, and Random Forest — trained on curated symptom-disease datasets. Preprocessing includes binary symptom encoding, StandardScaler normalization, mutual information-based feature selection, and PCA dimensionality reduction retaining 95% variance. The best-performing model is automatically selected based on test accuracy. Experimental evaluation achieved a classification accuracy of up to 95.1%, with sub-second prediction latency. The modular Streamlit-based interface supports five functional modules: disease prediction with confidence scoring, medicine recommendations with dosage and contraindication guidance, condition-specific diet plans, daily health routines, and an interactive analytics dashboard. The system is designed as an assistive tool to support informed preliminary health decisions.