Publication

Abstract : Chronic Fatigue Syndrome (CFS) is a complicated and exhausting condition that poses challenges for diagnosis because of its subjective symptoms and overlapping clinical signs. This study presents a hybrid model that combines Graph Neural Networks (GNNs) and XGBoost to predict CFS severity across seven outcome classes using longitudinal patient data from the University of Bristol CFS dataset. By capturing relational structures and high-dimensional features, the model achieved an accuracy of 97.77%, a precision of 97.83%, a recall of 97.77%, and an F1-score of 97.76%. These results demonstrate strong classification ability and clinical significance. Our method helps improve tracking of patient outcomes and supports the United Nations Sustainable Development Goals (SDGs), particularly SDG 3(Good Health and Well-being) and SDG 9(Industry, Innovation, and Infrastructure) by promoting accessibility of good health and well-being for all sections of society through AI-based disease prognosis. The use of AI in healthcare diagnostics shows a scalable and sustainable approach for managing chronic diseases, especially in areas with limited data or resources.