Grid search

Grid search is the exhaustive, systematic hyperparameter tuning technique: it evaluates every specified combination to find the optimal model configuration.

This method (also called a parameter sweep) defines a discrete grid across a model's hyperparameter space, then trains and validates the model for *every* possible combination. For instance, tuning a Random Forest might require testing three values for `n_estimators` and three values for `max_depth`, resulting in $3 \times 3 = 9$ total model fits. Performance is typically measured using k-fold cross-validation (e.g., $k=5$) to ensure robust results. While guaranteed to find the best combination within the defined grid, Grid Search is computationally expensive, especially as the number of hyperparameters or their value ranges increase.

https://www.w3schools.com/python/machine_learning_grid_search.asp