Extended Data Fig. 4: Machine learning (ML) trained models.

(a) Error plots for nine ML models using a 90%/10% training-test split. Training data points are represented by blue dots, while test data points are shown in red. The dashed line indicates where the predicted values match the experimental data (truth). (b) Root mean squared errors (RMSEs) depicting the prediction accuracy across the nine ML models trained on the dataset of 180 bioinspired hydrogels, assessed via 10-fold cross-validation. A lower test error, combined with minimized overfitting (i.e., a smaller gap between training and test errors), indicates a more effective regression model.