plot.csuv: Graphical illustration of selection uncertainty

In CSUV: Combined Selection and Uncertainty Visualiser (CSUV)

Description

Graphical illustration of selection uncertainty

Usage

## S3 method for class 'csuv'
plot(
  x,
  with.unconditional = FALSE,
  compare.method.fit = NULL,
  cv.mod = NULL,
  with.thr = TRUE,
  with.violin = FALSE,
  to.shade = TRUE,
  ci.method = "conditional",
  level = 0.1,
  var.freq.thr = 0.1,
  log.level = NULL,
  ...
)

Arguments

x	fitted results from CSUV::csuv()
with.unconditional	TRUE to get a unconditonal boxplot on the same graph. Default is FALSE
compare.method.fit	(optional) fitted results from CSUV::lm.compare.methods(). Alternatively, user can provide a data frame with each row contains the estimated coefficients from a method. The name of each row should be corresponding to the name of the method. The first value of each row should be the value of the intercept
cv.mod	(optional) a vector of estimated coefficients from cross validation. The first value should be the value of the intercept
with.thr	whether the selection by the CSUV should be show. Default is TRUE
with.violin	whether the graph with violin plot
to.shade	whether to shade the graph by the relative frequency calculated by CSUV. Default is TRUE
ci.method	how the confidence interval should be calculated. Default is "conditional"
level	the significant level of the whiskers. Default is 0.1
var.freq.thr	minimum variable frequency to show, default is 0.1
log.level	log level to set. Default is NULL, which means no change in log level. See the function CSUV::set.log.level for more details
...	additional argument for plot

Value

a ggplot object

Examples

X = matrix(rnorm(1000), nrow = 100)
Y = rowSums(X[,1:3])+rnorm(100)
mod.0 = csuv(X, Y, intercept = FALSE, q = 0, method.names = NULL)
cv.mod = lm.cv(X, Y, intercept = FALSE, fit.percent = 0.5, num.repeat = 50)
compare.mod = lm.compare.method(X, Y, intercept = FALSE)
plot(mod.0, compare.method.fit = compare.mod, cv.mod = cv.mod$est.b)

CSUV documentation built on Oct. 23, 2020, 5:49 p.m.