Skip to main content
Fig. 4 | BMC Medical Informatics and Decision Making

Fig. 4

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

Fig. 4

Ranking of feature importance and their correlations with the target variable (Short/Non-Short) for the best model in the Ensemble-Imbalance scenario. The x-axis on the left displays the average impact on the model prediction for each feature (mean absolute SHAP value). The x-axis on the right demonstrates the impact on the model prediction (SHAP value) concerning the feature values. Positive SHAP values indicate that the feature value led the model prediction toward the Short survival group, while negative SHAP values pushed it toward the Non-Short group

Back to article page