Documentation has incorrect default value for BayesSearchCV cross validation fold count.

[jgreenemi]

BaseSearchCV from scikit-learn sets cv to 5 if None is passed in, but the documentation for BayesSearchCV states that the default fold is 3 if the user doesn't specify a cv value on instantiation.

Since this is defined in scikit-learn and not within scikit-optimze itself, probably safer to omit the default number of folds altogether, or mention that this information can be found in the scikit-learn package, in case it changes again.