Fig. 5.

Identifying and tunning the importance of $x_i$ to the prediction of ${\mathrm{\Gamma }}_{pent}$. The (feature) importance of $x_i$ to $\alpha$ can be identified from 16 different configurations of $\underset{¯}{x}$, which can be further tuned by varying the additional flexibility factor, $n$ (Eq. 5). Interestingly, the optimal value of $n=-1$ is found to minimize the importance of $\underset{¯}{x}$, especially $x_3$ and $x_4$. This observation provides valuable insights into the choice of $\alpha$, as well as the establishment of the connection between ${\mathrm{\Gamma }}_{pent}$ and ${\mathrm{\Gamma }}_{2D}$.