Cloud-Based Early Detection of Parkinson’s Disease Using AI Tools in Amazon Web Services SageMaker

Abstract

Parkinson’s disease is a progressive neurological disorder that remains difficult to diagnose at early stages due to subtle symptom onset and reliance on subjective clinical assessments. Early identification is crucial to improving patient outcomes, reducing long-term healthcare costs, and enabling timely therapeutic interventions. This study proposes a cloud-based diagnostic framework leveraging Amazon Web Services (AWS) SageMaker to integrate diverse patient datasets including clinical records, speech samples, handwriting analysis, and wearable sensor data—for predictive modeling. The architecture incorporates an Extract, Transform, and Load (ETL) module for preprocessing, supervised machine learning models for training and validation, and a feedback loop for continuous refinement. Through interactive dashboards, healthcare professionals can access predictions and risk assessments in real time, supporting informed decision-making. Security and compliance are ensured through HIPAA-compliant encryption and role-based access control, while scalability is achieved via SageMaker’s managed infrastructure. Quantitative evaluation confirms high predictive sensitivity and accuracy, with heatmap and bar chart analyses identifying high-risk individuals and time-series tracking demonstrating effective progression monitoring, establishing the framework’s reliability in early Parkinson’s detection.