AI-Driven Heart Disease Prediction System Using Natural Language Processing and Machine Learning

Cardiovascular disease (CVD) is behind nearly 32 percent of deaths worldwide, which makes early detection essential. Recent work applies machine learning and deep learning methods, including Decision Trees, Random Forest, XGBoost, Gradient Boosting, SVM and hybrid neural networks, to predict heart disease from clinical attributes such as age, blood pressure, cholesterol and ECG readings, with reported accuracies ranging from 65 percent to above 98 percent. These systems, however, depend on structured clinical inputs that cannot be obtained without medical tests, which restricts their use in rural and resource-limited settings. This paper surveys recent studies on heart disease prediction and draws on their findings to design an AI-Driven Heart Disease Prediction System that removes this dependency. The proposed system uses natural language processing to pull symptoms out of plain-language descriptions and assigns a Low, Medium, or High risk level with Random Forest and Logistic Regression models, offered through a Streamlit interface so that non-medical users can carry out preventive screening.