predict.iregnet: Predict response using new covariates

In anujkhare/iregnet: Regularized interval regression

Description

Prediction is just X'beta for non-transformed distributions, and itrans(X'beta) for transformed distributions. In the case of gaussian, logistic and extreme_value, types link and response both return the linear predictors.

Usage

## S3 method for class 'iregnet'
predict(object, newx, lambda = NULL, type = c("link",
  "response"), ...)

Arguments

object	The S3 object of type iregnet returned by the iregnet method.
newx	New input matrix of covariates with dimension n_vars same as used for fit. Sparse matrices are not supported.
lambda	The values of lambda at which prediction is to be obtained. These should be a subset of the values on which the model was fit. To obtain predictions at other lambda values, re-fit the model. Default: NULL (object$lambda is used)
type	Type of prediction required. Type link returns the linear predictors. Type response returns the fitted values. In the case of gaussian, logistic and extreme_value, the link and response are the same.
...	Optional arguments. Currently unused.

anujkhare/iregnet documentation built on Aug. 23, 2019, 8:24 p.m.