Figure 6From: Multi-way metamodelling facilitates insight into the complex input-output maps of nonlinear dynamic modelsOptimalisation of the number of clusters in hierarchical metamodelling of the mammalian circadian clock model. A) Results from inverse hierarchical metamodelling using from 1–20 clusters in the N-way HC-PLSR. Left: Mean parameter prediction correlation coefficient (R2)-values within the calibration set, over the nine varied circadian clock model input parameters vs. the number of clusters used in the N-way HC-PLSR metamodelling. The calibration set observations were here treated as "new observations" (see Figure 2), and classified in the prediction stage. Using six clusters was considered optimal. Right: Parameter prediction R2-values within the calibration set for the nine different circadian clock model input parameters vs. the number of clusters used in the N-way HC-PLSR metamodelling. B) Results from classical hierarchical metamodelling using from 1–20 clusters in the N-way HC-PLSR. Left: Mean state variable prediction R2-values within the calibration set, over the 16 circadian clock model state variables vs. the number of clusters used in the N-way HC-PLSR metamodelling. The calibration set observations were here treated as "new observations", and classified in the prediction stage. Right: State variable prediction R2-values within the calibration set for the 16 circadian clock state variables vs. the number of clusters used in the N-way HC-PLSR metamodelling. Using six clusters was considered optimal.Back to article page