mod_lasso: Description of the function

In Marie-PerrotDockes/VariSel: Perform variable selection for different model types

Description

Description of the function

Description of the function

Value

The coefficients of the fused lasso ANCOVA for the different value of lambda

Super class

VariSel::VariSel -> mod_lasso

Methods

Public methods

• mod_lasso$new()

• mod_lasso$estime()

• mod_lasso$sel_cv()

• mod_lasso$sel_stabs()

• mod_lasso$plot_path()

• mod_lasso$clone()

Inherited methods

• VariSel::VariSel$get_coef()

• VariSel::VariSel$plot_anime()

• VariSel::VariSel$plot_BIC()

• VariSel::VariSel$plot_coef()

• VariSel::VariSel$plot_cv()

• VariSel::VariSel$plot_error()

• VariSel::VariSel$predict()

• VariSel::VariSel$ROC()

---

Method new()

Usage

mod_lasso$new(
  x,
  y,
  univ = TRUE,
  Sigma_12inv = diag(1, ncol(as.data.frame(y))),
  sepy = "__",
  penalty.factor = NULL
)

---

Method estime()

Usage

mod_lasso$estime(lambda = NULL, standardize = FALSE, intercept = TRUE)

Arguments

lambda

if the user wants to use it owns values of lambdas

---

Method sel_cv()

Usage

mod_lasso$sel_cv(s = "lambda.min")

---

Method sel_stabs()

Usage

mod_lasso$sel_stabs(nb.cores = 7, B = 500, PFER = 1)

---

Method plot_path()

Usage

mod_lasso$plot_path(type = "first", nb = 6)

---

Method clone()

The objects of this class are cloneable with this method.

Usage

mod_lasso$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

B <- c(1, -1, 1.5, 1.5, rep(0, 6), 2, 0, 2, 0)
group <- c(rep('M1', 10), rep('M2', 10))
regressors <- matrix(rnorm(6*20), ncol = 6)
X  <- model.matrix(~group + group:regressors - 1)
y <- X%*%B + rnorm(20)
y <- scale(y)
mod <- fl2(y, regressors, group)
colors <- c(rep("grey",2), rep('green',2),rep('black', 6), rep(c("orange","blue"), 2), 'darkgreen', rep('yellow',3), rep('purple',2))
matplot(mod$lambda ,t(mod$beta),type='l',col=colors)
#' @import R6 Matrix gglasso tidyverse glmnet stabs magrittr viridis stringr FusedLasso

Marie-PerrotDockes/VariSel documentation built on May 7, 2020, 1:09 a.m.