Health Prediction Based on Day-to-Day Life Activity using Machine Learning Approach

The potential of health surveillance systems to offer detailed insights into a person's overall well-being is now limited by their limits especially those that only analyze certain illnesses like Parkinson's disease. These approaches often ignore critical health indicators that are necessary for a comprehensive assessment of an individual's health, such as blood pressure, stress levels and heart health. This work proposes a complex framework that analyzes a broad variety of health data collected from smartwatch users to offer a deep understanding of various health factors. The system integrates several health parameters using state-of-the-art advanced data analysis and representation technologies to provide tailored suggestions based on historical trends, helping users make informed decisions for proactive health management. K-Means and Agglomerative Clustering algorithms are applied to group users based on their health characteristics after the data has been processed using the Standard Scaler and separated into testing and training sets. The optimal number of clusters is determined using the Elbow Method, with K-Means achieving a silhouette score of 0.85. The system's user-centric design facilitates the determination of areas for development and promotes a culture of preventative healthcare by offering extensive insights into health patterns. Ultimately, this approach seeks to empower users with practical information for better individual health management and societal well-being.