Deep Learning-Based Mobile Application for Cardiovascular Disease Prediction Integrating Environmental Factors

Abstract

Cardiovascular disease (CVD) remains a leading cause of death worldwide. Conventional machine-learning methods such as Support Vector Machines (SVM), K-Nearest Neighbors (KNN), and Logistic Regression have been widely used for CVD prediction, but their performance can vary across clinical datasets. In addition, many prediction systems do not incorporate environmental determinants such as air pollution. This study presents a deep-learning framework based on an Artificial Neural Network (ANN) and integrates real-time Air Quality Index (AQI) data into a mobile application. After preprocessing and comparative evaluation, the ANN trained on a 303-record clinical dataset achieved a test accuracy of 88.81%. A React Native application provides location-based AQI alerts, while a FastAPI backend handles prediction requests. The resulting system combines clinical prediction with environmental context to support accessible, real-time cardiovascular risk assessment.