Skip to main content
Fig. 6 | BMC Medical Informatics and Decision Making

Fig. 6

From: Ensemble-imbalance-based classification for amyotrophic lateral sclerosis prognostic prediction: identifying short-survival patients at diagnosis

Fig. 6

Examples of using the SHAP Decision plot to explain how the model classified patients into Short and Non-Short survival groups. Subfigures “a” and “b” show individualized classifications for each patient, where the process was conducted according to their feature values (displayed in gray font within parenthesis). Subfigure “c” shows the classification process comparing feature by feature for both patients. These graphs must be read from bottom to top. The slope of the line within each feature area indicates when the feature value drove the prediction toward the Non-Short (left sloping) or the Short (right sloping) groups. The longer the line length within the feature area, the more significant the impact of its value on the model prediction

Back to article page