Table 2.
R2 and root mean squared prediction error (RMSPE) corresponding to different regression models explaining the number of real active cases in the metropolitan area as a function of the natural logarithm of the viral load. RMSPE was obtained using a 6-fold cross validation procedure.
Model | R2 | RMSPE |
---|---|---|
Linear | 0.8515 | 581.94 |
GAM | 0.8767 | 508.62 |
LOESS (linear) | 0.8695 | 487.97 |
LOESS (quadratic) | 0.8833 | 478.33 |