matrixpls (version 0.7.0)

matrixpls.blindfold: Blindfold crossvalidation of predictions from matrixpls results

Description

matrixpls.blindfold Calculates blindfold predictions using matrixpls.

Usage

matrixpls.blindfold(data, ..., predictFun = stats::predict, nGroup = 4)

Arguments

data

Matrix or data frame containing the raw data.

...

All other arguments are passed through to matrixpls.

predictFun

The function used to calculate the predictions.

nGroup

The number of groups to divide the data into. Setting nGroup to the number of observations produces jackknife predictions.

Value

• A matrix of class matrixpls.blindfold containing predictions calculated with blindfolding.

Details

In blindfolding, the data are first divided into n equal sized groups. Then a statistic or a model is estimated using the data after omitting one of the groups. The results are then used to predict the observations in that group. The process is repeated when all groups have been predicted.