Build a set of models trained on different folds of cross-validated data. Can be used to estimate the uncertainty of a given model type at any point.
buildCVModel(x, y, control = list())design matrix (sample locations)
vector of observations at x
(list), with the options for the model building procedure:
types a character vector giving the data type of each variable. All but "factor" will be handled as numeric, "factor" (categorical) variables will be subject to the hamming distance.
target target values of the prediction, a vector of strings. Each string specifies a value to be predicted, e.g., "y" for mean, "s" for standard deviation.
This can also be changed after the model has been built, by manipulating the respective object$target value.
uncertaintyEstimator a character vector specifying which uncertaintyEstimator should be used.
"s" or the linearlyAdapted uncertrainty "sLinear". Default is "sLinear".
modellingFunction the model that shall be fitted to each data fold
set of models (class cvModel)