mbo_tramnet {tramnet} | R Documentation |
Uses model based optimization to find the optimal tuning parameter(s) in a regularized transformation model based on cross-validated log-likelihoods. Here the tramnet package makes use of the mlrMBO interface for Bayesian Optimization in machine learning problems to maximize the cv-logLik as a black-box function of the tuning parameters alpha and lambda.
mbo_tramnet(object, fold = 2, n_design = 5, n_iter = 5, minlambda = 0, maxlambda = 16, minalpha = 0, maxalpha = 1, folds = NULL, learner = "regr.km", pred.type = "se", opt_crit = makeMBOInfillCritEI(), noisy = FALSE, obj_type = c("lasso", "ridge", "elnet"), verbose = TRUE, ...)
object |
object of class |
fold |
fold for cross validation |
n_design |
results in |
n_iter |
number of iterations in the model based optimization procedure |
minlambda |
minimum value for lambda (default: |
maxlambda |
maximum value for lambda (default: |
minalpha |
minimum value for alpha (default: |
maxalpha |
maximum value for alpha (default: |
folds |
self specified folds for cross validation (mainly for reproducibility and comparability purposes) |
learner |
type of leaner used for the optimization (default: |
pred.type |
prediction type of the learner (default: |
opt_crit |
optimization criterion, default: expected improvement |
noisy |
indicates whether folds for k-fold cross-validation should
be random for each iteration, leading to a noisy objective function
(default: |
obj_type |
objective type, one of |
verbose |
toggle for a verbose output (default: |
... |
additional arguments are ignored |
returns an object of class "MBOSingleObjResult"
which is
documented in mbo
set.seed(24101968) data("Prostate", package = "lasso2") x <- scale(model.matrix(lpsa ~ . - 1, data = Prostate)) y <- Prostate$lpsa m0 <- BoxCox(lpsa ~ 1, data = Prostate) mt <- tramnet(m0, x = x, lambda = 0, alpha = 1, check_dcp = FALSE) (run_lasso <- mbo_tramnet(object = mt, n_design = 2, n_iter = 2, obj_type = "lasso"))