What does “trim” mean in “apply()”?

Question

What does the argument "trim" mean in "apply()"?

For example, I would like to calculate the mean of each column of an array x :

apply(x, 2, mean, trim = .2)

How is it different from

apply(x, 2, mean)

Which one shall I use?

If I would like to calculate the standard deviation of each column of an array x , which one shall I use

apply(x, 2, sd, trim = .2)

or

apply(x, 2, sd)

Thanks!

Answer 1

The trim argument is not to do with apply , but to do with mean .

From ?mean :

trim : the fraction (0 to 0.5) of observations to be trimmed from each end of x before the mean is computed. Values of trim outside that range are taken as the nearest endpoint.

sd doesn't have an argument named trim , so if you tried apply(x, 2, sd, trim = .2) you would get an error about trying to use an unused argument.

---

This is not an argument to apply , but it is an argument passed down from apply to mean using ... :

From ?apply :

... : optional arguments to FUN .

This adds flexibility and reduces dependence on creating anonymous functions. For example, without the ... argument, if you wanted to do a trimmed mean, you would have to do:

apply(x, 2, function(x) mean(x, trim = .2))

Answer 2

Just tried this with dplyr and it seems to work fine.

df <- x %>%
    select(V1:V7) %>%
    summarise_each(funs(mean(., na.rm=T, trim=0.2)))