Definition
A hyperparameter optimization method that randomly samples parameter combinations from a specified distribution, often proving more efficient than Grid Search.
In short: A hyperparameter optimization method that randomly samples parameter combinations from a specified distribution, often proving more efficient than Grid Search.