Hello,
Try the following.
out <- lapply(df, function(x){
qq <- quantile(x, probs = 0.99)
which(x > 5*qq)
})
out
The list 'out' contains indices to the outliers. You can now have those
outliers as follows
df[out[[1]], 1] # first column
Etc.
Hope this helps,
Rui Barradas
I have a data frame with some extreme values which I wish to identify
and repeat an analysis without these extreme values. How could I
identify several columns with values which are 5 times higher than the
99th percentile?
Sample data is pasted below.
> dput(df)
structure(list(ad1 = c(98, 6.9, 8
2 matches
Mail list logo
