CROSSVAL

Error/performance estimation by cross validation (rotation)

    [ERR,CERR,NLAB_OUT] = CROSSVAL(A,CLASSF,NFOLDS,1,TESTFUN)
    [ERR,STDS] = CROSSVAL(A,CLASSF,NFOLDS,NREP,TESTFUN)
    [ERR,CERR,NLAB_OUT] = CROSSVAL(A,CLASSF,NFOLDS,'DPS',TESTFUN)
    R = CROSSVAL(A,[],NFOLDS,0)

Input
A	Input dataset
CLASSF	The untrained classifier to be tested.
NFOLDS	Number of folds  (default: number of samples: leave-one-out)
NREP	Number of repetitions (default: 1)
TESTFUN	Mapping,evaluation function (default classification error)

Output
ERR	Average test error or performance weighted by class priors.
CERR	Unweighted test errors or performances per class
NLAB_OUT	Assigned numeric labels
STDS	Standard deviation over the repetitions.
R	Index array with rotation set

Description

Cross validation estimation of the error or performance (defined by TESTFUN) of the untrained classifier CLASSF using the dataset A. The set is randomly  permutated and divided in NFOLDS (almost) equally sized parts, using a  stratified procedure. The classifier is trained on NFOLDS-1 parts and the  remaining part is used for testing. This is rotated over all parts. ERR is  their weighted avarage over the class priors. CERR are the class error  frequencies. The inputs A and/or CLASSF may be cell arrays of datasets and  classifiers. In that case ERR is an array with on position ERR(i,j) the  error or performance of classifier j for dataset i. In this mode CERR and  NLAB_OUT are returned in cell arrays.

For NREP > 1 the mean error(s) over the repetitions is returned in ERR and the standard deviations in the observed errors in STDS.

If NREP == 'DPS', crossvalidation is done by density preserving data  splitting (DPS). In this case NFOLD should be a power of 2.

In case NREP == 0 an index array is returned pointing to a fold for every  object. No training or testing is done. This is useful for handling  training and testing outside CROSSVAL.

Note that this routine is identical to the PRCROSSVAL rouitne located in  the PRTools main directory. It thereby avoids confusion with the CROSSVAL routine in the Stats toolbox if called by a dataset as a first parameter.  Users should preferably call PRCROSSVAL for the PRTools routine and use  CROSSVAL for the Stats version.

Reference(s)

1. R. Kohavi: A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection. IJCAI 1995: 1137-1145.
2. M. Budka, B. Gabrys, Correntropy-based density-preserving data sampling as an alternative to standard cross-validation, IJCNN2010, 1-8

See also

datasets, mappings, dps, cleval, testc, prcrossval,