ProSGPV (version 1.0.0)

plot.sgpv: plot.sgpv: Plot variable selection results

Description

S3 method plot for an object of class sgpv. When the two-stage algorithm is used, this function plots the fully relaxed lasso solution path on the standardized scale and the final variable selection results. When the one-stage algorithm is used, a histogram of all coefficients with selected effects is shown.

Usage

# S3 method for sgpv
plot(x, lpv = 3, lambda.max = NULL, short.label = T, ...)

Arguments

x

An sgpv object

lpv

Lines per variable. It can take the value of 1 meaning that only the bound that is closest to the null will be plotted, or the value of 3 meaning that point estimates as well as 95% confidence interval will be plotted. Default is 3.

lambda.max

The maximum lambda on the plot. Default is NULL.

short.label

An indicator if a short label is used for each variable for better visualization. Default is TRUE

...

Other plot arguments

Examples

# NOT RUN {
# prepare the data
x <- t.housing[, -ncol(t.housing)]
y <- t.housing$V9

# one-stage algorithm
out.sgpv.1 <- pro.sgpv(x = x, y = y, stage = 1)

# plot the selection result

plot(out.sgpv.1)

# two-stage algorithm
out.sgpv.2 <- pro.sgpv(x = x, y = y)

# plot the fully relaxed lasso solution path and final solution
plot(out.sgpv.2)

# zoom in a little bit
plot(out.sgpv.2, lambda.max = 0.01)

# only plot one confidence bound
plot(out.sgpv.2, lpv = 1, lambda.max = 0.01)
# }