Figure 12From: Machine learning applied to simulations of collisions between rotating, differentiated planetsPerformance as a function of training set size. Performance on the holdout test dataset (quantified by \(r^{2}\)-scores) is shown as a function of training set size (TSS). Regression performance for a well-performing parameter \(M_{\mathrm{LR}}\) is shown in the top panel and a relatively difficult to regress parameter \(\theta _{\mathrm{SLR}}\) the lower panelBack to article page