Fig. 4: Importance and contribution of the explanatory factors to the established model for weekly new admissions prediction.
a The importance of explanatory factors was ranked by the increase in %MSE (percent change in mean square error when the explanatory factor is permuted). A higher increase in %MSE corresponds to higher importance. The significance of the explanatory factors was marked as *, **, and *** representing a p value of ≥0.01 and <0.05, ≥0.001 and <0.01 and <0.001, respectively. Hos1w, Hos2w, Hos3w, Hos4w are the upcoming week, the second, third and fourth week after the wastewater sampling, respectively. The two-factor partial dependence for predicting weekly new admissions at the second week (Hos2w, subfigure b) and fourth week (Hos4w, subfigure c), on CRNA and four significant explanatory factors used in the models. The horizontal axis represents the values of CRNA, whereas the vertical axis represents the values of the other four explanatory factors (as shown in the title). The color gradients in the figure indicate the partial dependence of the predicted target concerning a specific x-value and y-value combination.
