Using: the package "glmnet"

Problem: I use the plot function to plot a lasso image, I feel the labels are too small. So, I want to change the cex, but, it's not working. I looked up the documents of "glmnet", the plot function seems like normal plot. Any idea?

My code:

plot(f1, xvar='lambda', label=TRUE, cex=1.5)

Other: I tried like cex.lab=3. It worked but for the x-y axis labels.

CodePudding user response：

What you are actually using is the generic plot which dispatches a method depending on the class of the object. In this case,

class(fit1)
# [1] "elnet"  "glmnet" 

glmnet:::plot.glmnet will be used, which internally uses glmnet:::plotCoef. And here lies the problem; in glmnet:::plotCoef the respective cex parameter in glmnet:::plotCoef is hard coded to 0.5.―We need a hack:

plot.glmnet <- function(x, xvar=c("norm", "lambda", "dev"), 
          label=FALSE, ...) {
  xvar <- match.arg(xvar)
  plotCoef2(x$beta, lambda=x$lambda, df=x$df, dev=x$dev.ratio,        ## changed
           label=label, xvar=xvar, ...)
}

plotCoef2 <- function(beta, norm, lambda, df, dev, label=FALSE, 
          xvar=c("norm", "lambda", "dev"), xlab=iname, ylab="Coefficients",
          lab.cex=0.5, xadj=0, ...) {
  which <- glmnet:::nonzeroCoef(beta)
  nwhich <- length(which)
  switch(nwhich   1, `0`={
    warning("No plot produced since all coefficients zero")
    return()
  }, `1`=warning("1 or less nonzero coefficients; glmnet plot is not meaningful"))
  beta <- as.matrix(beta[which, , drop=FALSE])
  xvar <- match.arg(xvar)
  switch(xvar, norm={
    index=if (missing(norm)) apply(abs(beta), 2, sum) else norm
    iname="L1 Norm"
    approx.f=1
  }, lambda={
    index=log(lambda)
    iname="Log Lambda"
    approx.f=0
  }, dev={
    index=dev
    iname="Fraction Deviance Explained"
    approx.f=1
  })
  dotlist <- list(...)
  type <- dotlist$type
  if (is.null(type)) 
    matplot(index, t(beta), lty=1, xlab=xlab, ylab=ylab, type="l", ...)
  else matplot(index, t(beta), lty=1, xlab=xlab, ylab=ylab, ...)
  atdf <- pretty(index)
  prettydf <- approx(x=index, y=df, xout=atdf, rule=2, method="constant", f=approx.f)$y
  axis(3, at=atdf, labels=prettydf, tcl=NA)
  if (label) {
    nnz <- length(which)
    xpos <- max(index)
    pos <- 4
    if (xvar == "lambda") {
      xpos <- min(index)
      pos <- 2
    }
    xpos <- rep(xpos   xadj, nnz)                         ## changed
    ypos <- beta[, ncol(beta)]
    text(xpos, ypos, paste(which), cex=lab.cex, pos=pos)  ## changed
  }
}

If we load the two hacked functions, we now can adapt the coefficient labels. New are the parameters lab.cex for size and xadj to adjust the x position of the numbers.

plot(fit1, xvar='lambda', label=TRUE, lab.cex=.8, xadj=.085)

Data:

set.seed(122873)
x <- matrix(rnorm(100 * 10), 100, 10)
y <- rnorm(100)
fit1 <- glmnet(x, y)