cv.sprinter: Running sprinter with cross-validation
In hugogogo/sprintr: Sparse Reluctant Interaction Modeling

Description Usage Arguments Value See Also Examples

The main cross-validation function to select the best sprinter fit for a path of tuning parameters.

cv.sprinter(
  x,
  y,
  square = FALSE,
  num_keep = NULL,
  lambda1 = NULL,
  lambda3 = NULL,
  cv_step1 = FALSE,
  nlam1 = 10,
  nlam3 = 100,
  lam_min_ratio = ifelse(nrow(x) < ncol(x), 0.01, 1e-04),
  nfold = 5,
  foldid = NULL,
  verbose = FALSE,
  ...
)

`x`	An `n` by `p` design matrix of main effects. Each row is an observation of `p` main effects.
`y`	A response vector of size `n`.
`square`	Indicator of whether squared effects should be fitted in Step 1. Default to be FALSE.
`num_keep`	A user specified number of candidate interactions to keep in Step 2. If `num_keep` is not specified (as default), it will be set to `round[n / log n]`.
`lambda1`	Tuning parameter values for Step 1. `lambda1` is a vector. Default to be NULL, and the program will compute its own `lambda1` based on `nlam1` and `lam_min_ratio`.
`lambda3`	Tuning parameter values for Step 3. `lambda3` is a matrix, where the k-th column is the list of tuning parameter in Step 3 corresponding to Step 1 using `lambda1[k]`. Default to be NULL, and the program will compute its own `lambda3` based on `nlam3` and `lam_min_ratio`.
`cv_step1`	Indicator of whether cross-validation of `lambda1` should be carried out in Step 1 before subsequent steps. Default is `FALSE`.
`nlam1`	the number of values in `lambda1`. If not specified, they will be all set to `10`.
`nlam3`	the number of values in each column of `lambda3`. If not specified, they will be all set to `100`.
`lam_min_ratio`	The ratio of the smallest and the largest values in `lambda1` and each column of `lambda2`. The largest value is usually the smallest value for which all coefficients are set to zero. Default to be `1e-2` in the `n` < `p` setting.
`nfold`	Number of folds in cross-validation. Default value is 5. If each fold gets too view observation, a warning is thrown and the minimal `nfold = 3` is used.
`foldid`	A vector of length `n` representing which fold each observation belongs to. Default to be `NULL`, and the program will generate its own randomly.
`verbose`	If `TRUE`, a progress bar shows the progress of the fitting.
`...`	other arguments to be passed to the `glmnet` calls, such as `alpha` or `penalty.factor`

An object of S3 class "sprinter".

n: The sample size.
p: The number of main effects.
square: The square parameter passed into sprinter.
a0_step3: Estimate of intercept corresponding to the CV-selected model.
compact: A compact representation of the selected variables. compact has three columns, with the first two columns representing the indices of a selected variable (main effects with first index = 0), and the last column representing the estimate of coefficients.
fit: The whole glmnet fit object.
fitted: fitted value of response corresponding to the CV-selected model.
num_keep: The value of num_keep.
cvm: The averaged estimated prediction error on the test sets over K folds.
cvse: The standard error of the estimated prediction error on the test sets over K folds.
foldid: Fold assignment. A vector of length n.
i_lambda1_best: The index in lambda1 that is chosen by CV by minimizing cvm.
i_lambda3_best: The index in lambda3 that is chosen by CV by minimizing cvm.
lambda1_best: The value of lambda1 that is chosen by CV by minimizing cvm.
lambda3_best: The value of lambda3 that is chosen by CV by minimizing cvm.
call: Function call.

predict.cv.sprinter

n <- 100
p <- 100
x <- matrix(rnorm(n * p), n, p)
y <- x[, 1] - 2 * x[, 2] + 3 * x[, 1] * x[, 3] - 4 * x[, 4] * x[, 5] + rnorm(n)
mod <- cv.sprinter(x = x, y = y)