crossvalidator crossvalidation class.
 
    DESCRIPTION
    The crossvalidator class performs crossvalidation. It uses a
    particular crossvalidation on a multivariate analysis specified in the
    property mva.
 
    One may use 'type' equal to:
    - 'nfold' : n-fold cross-validation specified by 'folds' (default: 5)
    - 'split' : data is split into train and test data; amount of training
                data is given by 'proportion' (default: 0.75)
    - 'loo'   : leave-one-out cross-validation; takes 1 trial per fold
    - 'bloo'  : balanced loo cv; takes 1 trial from each class per fold.
                Assumes that each class has the same nr of trials.
   
    Alternatively one can specify the cell-arrays 'trainfolds' and
    'testfolds' to contain for each fold the trial indices that need to be
    used. If one of these is specified then the other one is automatically
    filled using the complement of the trials. If these cell-arrays are
    non-empty then the settings above are ignored.
 
    In order to balance the occurrence of different classes one may set
    'resample' equal to true (default: false). Resample will upsample less
    occurring classes during training and downsample often occurring
    classes during testing.
 
    In case of memory problems, use 'compact' equal to true (default:
    false). This will not store the trained multivariate methods but just
    the classification outcomes and the parameters that are returned by the
    model function of the multivariate methods.
 
    EXAMPLE
    X = rand(10,20); Y = [1 1 1 1 1 2 2 2 2 2]';
    m = dml.crossvalidator('mva',dml.svm)
    m = m.train(X,Y);
    m.statistic('accuracy')
 
    DEVELOPER
    Marcel van Gerven (m.vangerven@donders.ru.nl)