A global optimization strategy used to find the minimum or maximum of an objective function that is expensive to evaluate, applied here to tune machine learning models for optimal performance.