CODEWITHSUNDEEP
r
An old man in the sea.
An old man in the sea.

Reputation: 1508

NaN error in function in R

The objective of my function is to calculated a truncated mean, i.e., you give a sample, and then compute the mean for all the values between certain quantiles.

To do that, I've written

 meank<-function(x,k){
  xt<-quantile(x,c(k,1-k));
  mean(x[x>xt[1] & x < xt[2] ])
}

When doing meank(rnorm(100),0.4), it returns NaN. Why is that?

Any help would be appreciated.

Upvotes: 0

Views: 1929

Answers (1)

Dilettant
Dilettant

Reputation: 3335

In the implementation of meank you cancel all values outside of [k, 1-k]"percent" of range, thus when no value is left in, the subsequent meanhas to yield NaN.

Reproducible test:

> meank(c(1, 2, 2),0.4)
[1] NaN

Details step wise evaluation:

> quantile(c(1, 2, 2),c(0.4,1-0.4))
40% 60% 
1.8 2.0

and:

> xt <- quantile(c(1, 2, 2),c(0.4,1-0.4))
> mean(c(1, 2, 2)[c(1, 2, 2)>xt[1] & c(1, 2, 2) < xt[2] ])
[1] NaN

Upvotes: 1

Related Questions