Random Forest with Times Series
Yigit Aydede · 2023
In this chapter, we turn our attention to the application of embedding techniques for direct forecasting using Random Forests. The choice to employ Random Forests is driven by its inherent advantages, such as not requiring explicit tuning through grid search. Nevertheless, in practice, we can still optimize the model by searching for the optimal number of trees and the number of variables randomly sampled as candidates at each split.