Fig. 3
From: An interpretable machine learning-assisted diagnostic model for Kawasaki disease in children
Global model explanation. (a) SHAP summary bar plot. This plot displays the ranking of feature importance based on Shapley values. (b) SHAP summary point plot. Each point represents a Shapley value for a specific patient and feature. Color of points represents actual feature values, where red represents higher values and blue represents lower values. The horizontal-axis displays the Shapley values, while the vertical-axis depicts features ordered by their mean absolute Shapley values. (c) SHAP dependence plot. Each individual subplot illustrating the impact of a single feature on the model’s prediction. Each point corresponds to a single patient, with SHAP values on the vertical-axis and actual feature values on the horizontal-axis.
