f.fit: fit a spline to the histogram of imp

In RFlocalfdr: Significance Level for Random Forest Impurity Importance Scores

fit a spline to the histogram of imp

Description

fit a spline to the histogram of imp

Usage

f.fit(imp, df = 10, debug.flag = 0, temp.dir = NULL)

Arguments

imp	the variable importances
df	the degrees of freedom for the spline fit
debug.flag	either 0 (no debugging information), 1 or 2
temp.dir	if debug flag is >0 then information is written to temp.dir

Value

a list with the following components

• "x" – midpoints of the histogram

• "zh" – a histogram object as returned by "hist"

• "f.spline" – the spline fit. The fit is given by a glm mode glm(zh$counts ~ splines::ns(x), poisson)

• "counts" the counts from the histogram

Examples

data(imp20000)
imp <- log(imp20000$importances)
res <- f.fit(imp)
plot(res$zh, xlab="importances", main="histogram of importances")
points(res$midpoints,res$counts, col="grey90")
lines(res$zh$breaks[-1],res$f.spline,col="blue", lwd=3)
legend("topleft",c("spline fit"), col="blue", lwd=3)

RFlocalfdr documentation built on April 4, 2025, 5:16 a.m.