There's more...

In cross-validation, all available data is used, in groups of a fixed size, alternatively as a testing and as a training set. Therefore, each pattern is either classified (at least once) or used for training. The performances obtained depend, however, on the particular division. Therefore, it may be useful to repeat cross-validation several times in order to become independent of the particular division.