Definition
A hyperparameter optimization method that samples random combinations of parameter values from specified distributions.
Detailed Explanation
Random search randomly samples parameter values from defined distributions rather than testing all combinations. Research has shown it to be more efficient than grid search, especially when only a few hyperparameters are important. It can find good parameter settings with fewer trials than grid search.
Use Cases
Hyperparameter optimization, Model tuning, Efficient parameter search