Fig. 6: Model performance and feature analysis after feature engineering. | npj Computational Materials

Fig. 6: Model performance and feature analysis after feature engineering.

From: A machine learning model with minimize feature parameters for multi-type hydrogen evolution catalyst prediction

Fig. 6

a Accuracy of the ETR model on the training and testing sets with ten features b The SHAP value distribution of feature φ. c Global interpretation (average feature importance) and local interpretation (SHAP value distribution) and d SHAP values of each input feature importance on the ETR model with ten features \(\Delta {G}_{H}\). The contribution of each feature to \({\Delta G}_{H}\) predictions for e ReN₂ and f Ru₃Pb. g The interaction plot for all features based on importance.

Back to article page