This invention relates to the field of
medical information processing technology, and more particularly to a method,
system, and platform for predicting
gamma globulin-unresponsive
Kawasaki disease. Clinical data from affected children is collected, preprocessed, and feature-selected to obtain eight key features. An
ensemble learning algorithm is used to construct and validate a prediction model. During prediction, the eight features are input into the model, and the output
risk probability is used. The results are then interpreted using SHAP (Shape
Algorithm for Predicting Risk) to generate a visual representation. The
system is deployed as a prediction platform via a web framework. This invention is based on multi-center data, exhibits strong generalization ability and good
interpretability, and is easy to apply rapidly in clinical practice.