Cross-validate an elastic net — elastic_net

glmnet's cv.glmnet() on the Rust core: on each fold, fits the whole lambda path to the other folds (warm starts) and scores the family's mean deviance on the held-out fold; folds run in parallel. The score per lambda is the folds' mean weighted by fold weight, with its standard error. Matches cv.glmnet() on the same folds.

Usage

elastic_net_cv(
  formula,
  data,
  family = "poisson",
  link = NULL,
  alpha = 1,
  lambda = NULL,
  nlambda = 100,
  lambda_min_ratio = 1e-04,
  standardize = TRUE,
  penalty_factor = NULL,
  offset = NULL,
  weights = NULL,
  folds = 10,
  seed = 1,
  foldid = NULL,
  theta = NULL,
  power = NULL,
  link_power = NULL
)

Arguments

formula: A model formula; offset(...) terms are honoured.
data: A data frame.
family: "gaussian", "poisson", "gamma", "inverse_gaussian", "binomial" (response a proportion, weights the trials), "negative_binomial" (needs theta) or "tweedie" (needs power).
link: "identity", "log", "logit", "probit", "cloglog", "inverse", "inverse_squared" or "power" (needs link_power); NULL for the family's canonical link.
alpha: Mixing between ridge (0) and the lasso (1).
lambda: Penalty strengths; NULL for a path of nlambda values log-spaced from the smallest lambda that zeroes every coefficient down to lambda_min_ratio times it.
nlambda, lambda_min_ratio: The default path.
standardize: Penalize standardized coefficients.
penalty_factor: Optional penalty factors: a vector with one per design column, or a named vector for some columns (the rest get 1); 0 leaves a column unpenalized.
offset: Optional offset added to any offset() terms.
weights: Optional prior weights.
folds: Number of folds, used when foldid is NULL.
seed: Seed of the fold assignment.
foldid: Optional fold number per row.
theta: Negative binomial theta (variance mu + mu^2 / theta).
power: Tweedie power in (1, 2).
link_power: Exponent of the power link.

Value

A list with lambda, mean, se, lambda_min (lowest mean) and lambda_1se (the largest lambda within one standard error of it), and fit, the elastic_net_model on all rows over lambda.

Examples

set.seed(1)
d <- data.frame(x1 = rnorm(80), x2 = rnorm(80))
d$y <- 1 + 2 * d$x1 + rnorm(80)
cv <- elastic_net_cv(y ~ x1 + x2, d, family = "gaussian", nlambda = 30)
coef(cv$fit, lambda = cv$lambda_1se)
#> (Intercept)          x1          x2 
#>    1.083230    1.744625    0.000000