plot.l4acp: plot the threshold grid curve produced by l4acp

In wsggong/toronto: The Lasso for High Dimensional Regression with a Possible Change Point

Description

Plots the lasso objective function's value obtained from running regression by putting each value of threshold parameter's range in a model. It can also make plot not from lasso objective function but from ordinary MSE. Since l4acp can use multiple lambda values, it is required to specify a lambda value. Note that the plot has x-axis corresponding to threshold grid, and y-axis corresponding to the lasso objective function.

Usage

## S3 method for class 'l4acp'
plot(x, v = 1, type = "lasso", ...)

Arguments

x	fitted "l4acp" object
v	inex for a lambda value. Remember that you can only make a plot with one lambda value. By default it is 1, which is the largest among the lambda sequence used for l4acp. It is consistent to the column names of a matrix that comes out when you call the fitted model in console.
type	If the type is "lasso", it makes the plot by the value of a lasso objective function. Else if the type is "mse", it makes the plot by the value of a MSE.
...	other arguments that can be applied to l4acp

Details

A plot is produced, and nothing is returned

Examples

# using 'lr' as an threshold variable
data("grdata")
lrdata <- grdata[(!is.na(grdata[,"lr"])),]
x <- lrdata[,c(4:dim(grdata)[2])]
q <- lrdata[,"lr"]
y <- lrdata[,"gr"]
fit <- l4acp(x=x, y=y, q=q, s=0.00094)
plot(fit, type="lasso")

wsggong/toronto documentation built on May 15, 2019, 1:21 p.m.