In machine learning, randomized search is a hyperparameter optimization method which takes a distribution of parameter combinations, trains a model using a random sampling of those combinations, and then returns the combination that best optimizes a specified evaluation metric.
