cv.wtf1: Monotone Degree-One Weighted Trend Filter Cross-Validation

In barbehenna/coleReg: COmpound Loss Emporium with Regression

Monotone Degree-One Weighted Trend Filter Cross-Validation

Description

Use a fixed set of knots and a random partition of the data to cross-validate the risk estimate for a provided list of tau values. If 'nfolds = length(x)' leave-one-out cross-validation is used. If 'fit.best = TRUE' (default) then a final model is fit using all of the data and best value of tau. A seed is available to set for reproducible partitions.

Usage

cv.wtf1(x, s, tau.options, nfolds = 5, seed = NULL, fit.best = TRUE, ...)

Arguments

x	Gaussian sequence
s	standard deviation
tau.options	a vector of candidate values for TV(d')
nfolds	number of cross-validation folds
seed	random seed used for reproducible folds
fit.best	boolean (default TRUE) whether to fit all of 'x' with the best candidate value of TV(d')
...	additional parameters passed to wtf1() such as backend and knots

Value

theta_hat	estimated values of means of Gaussian sequence
x	original Gaussian sequence
s	known standard deviation
SURE	value of minimum risk estimate
tau	the user specified bound on TV(d')
TV1	the value of TV(d') for minimizer
intercept	intercept of fitted estimator
slopes	slopes of fitted estimator
knots	location of knots
backend	which version of wtf1 was used
solver	solver used by CVXR
...	additional inputed parameters
cv.risks	data.frame containing cross-validation parameters and risks
nfold	number of folds for cross-validation
seed	seed for reproducible fold partitions

Examples

# basic usage
set.seed(1)
theta = rnorm(250)
x = theta + rnorm(250)
res = cv.wtf1(x, s = 1, tau.options = 1:2, knots = -2:2)
mean((theta - res$theta_hat)^2)

barbehenna/coleReg documentation built on May 8, 2022, 12:05 a.m.