Predict responses for observations not used for training

returns
cross-validated predicted responses by the cross-validated linear
regression model `YHat`

= kfoldPredict(`CVMdl`

)`CVMdl`

. That is, for every fold, `kfoldPredict`

predicts
responses for observations that it holds out when it trains using
all other observations.

`YHat`

contains predicted responses for each
regularization strength in the linear regression models that compose `CVMdl`

.

`fitrlinear`

| `predict`

| `RegressionLinear`

| `RegressionPartitionedLinear`