Model validation was developed using an independent dataset of 40 new windows. The obtained model was then applied to the new data, and the Pearson coefficient and root-mean-square error (RMSE) were used to measure the predictive capacity of the model.