Fig. 9: Boxplots for performance comparison of the three models under spatial and random cross-validation. | npj Heritage Science

Fig. 9: Boxplots for performance comparison of the three models under spatial and random cross-validation.

From: Machine-learning-based reconstruction of Ming-dynasty defensive corridors in Yuxian

Fig. 9: Boxplots for performance comparison of the three models under spatial and random cross-validation.

a–c correspond to the boxplots of AUC, PR-AUC, and Brier score, respectively. The purple color represents the 7 km spatial block CV, and the yellow color represents the random stratified CV. The boxes and whiskers indicate the distribution of results across the five folds, while the triangles denote the inter-fold mean values. These are used to compare the discriminative ability and stability of different models and various CV schemes.

Back to article page