Figure 1
Panel (a): The heatmap titled “Model Comparison for R M S E” shows a square grid where both the horizontal axis and vertical axis list models in the same order: “r f 100”, “r f 300”, “s v r”, “l i g h t g b m”, “x g b o o s t”, “o l s”, “k n n”, “Decision Tree”, and “A N N”. Each cell represents a pairwise comparison between the model on the horizontal axis and the model on the vertical axis. The diagonal cells represent self-comparison. The grid shows darker shaded cells concentrated in comparisons involving “s v r”, “o l s”, “k n n”, “Decision Tree”, and “A N N”, indicating stronger relative performance patterns, while lighter shaded cells appear more frequently in comparisons involving “r f 100”, “r f 300”, and “l i g h t g b m”. The pattern is symmetric across the diagonal. Panel (b): The heatmap titled “Model Comparison for R-squared” shows a square grid where both the horizontal axis and vertical axis list models in the following order: “r f 100”, “r f 300”, “s v r”, “l i g h t g b m”, “x g b o o s t”, “o l s”, “k n n”, “Decision Tree”, and “A N N”. Each cell represents a pairwise comparison between models. The diagonal represents self-comparison. Darker shaded cells appear more prominently across comparisons involving “s v r”, “k n n”, “Decision Tree”, and “A N N”, while lighter shading is visible around “o l s” and “x g b o o s t” in some comparisons. The grid maintains symmetry across the diagonal. Panel (c): The heatmap titled “Model Comparison for M A E” shows a square grid where both the horizontal axis and vertical axis list models in the following order: “r f 100”, “r f 300”, “s v r”, “l i g h t g b m”, “x g b o o s t”, “o l s”, “k n n”, “Decision Tree”, “A N N”. Each cell represents a pairwise comparison between models, and diagonal cells represent self-comparison. Darker shaded cells appear across comparisons involving “s v r”, “o l s”, “k n n”, “Decision Tree”, and “A N N”, while lighter shaded cells appear more frequently in comparisons involving “r f100”, “r f300”, and “x g b o o s t”. The pattern remains symmetric across the diagonal.Comparison of performance across machine learning models. Source: Authors’ own work