Use a `cv_coxnet` Result

Predict from, tidy, plot, or collect the metrics of a [cv_coxnet()] result. Predictions and coefficients come from the model refit to all the data at the chosen `mixture`.

Usage

# S3 method for class 'cv_coxnet'
predict(
  object,
  new_data,
  type = c("linear_pred", "survival"),
  penalty = "lambda.min",
  eval_time = NULL,
  increasing = TRUE,
  ...
)

# S3 method for class 'cv_coxnet'
tidy(x, penalty = "lambda.min", ...)

# S3 method for class 'cv_coxnet'
collect_metrics(x, ..., summarize = TRUE)

# S3 method for class 'cv_coxnet'
autoplot(object, ...)

Arguments

object, x

A [cv_coxnet()] result.

new_data

A data frame of new predictors.

type

`"linear_pred"` (the default), `"survival"`, or `"time"` (restricted mean survival time), as for [predict.coxnet_model()].

penalty

`"lambda.min"` (default), `"lambda.1se"`, or a number.

eval_time, increasing

As for [predict.coxnet_model()].

...

Not used.

summarize

For `collect_metrics()`: `TRUE` for means over resamples, `FALSE` for each resample's metrics.

Value

`predict()`: a tibble as from [predict.coxnet_model()]. `tidy()`: the coefficients at `penalty`. `collect_metrics()`: a tibble of metrics. `autoplot()`: a ggplot of the selection metric against the penalty, with `lambda.min` (solid) and `lambda.1se` (dashed) marked.