I am trying to use the "kernel" function. To understand how it works I tried out some of the examples.
None of them works as shown in the following:
> kernel("daniell", 50) #
Error in kernel("daniell", 50) : unused argument(s) (50)
> kernel("daniell", 10) #
Error in kernel("daniell", 10) : unused argument(s) (10)
> kernel("daniell", c(3,3))
Error in kernel("daniell", c(3, 3)) : unused argument(s) (c(3, 3))
> kernel("fejer", 100, r=6)
Error in kernel("fejer", 100, r = 6) : unused argument(s) (100, r = 6)
> kernel("daniell", c(11,7,3) )
Error in kernel("daniell", c(11, 7, 3)) :
unused argument(s) (c(11, 7, 3))
My goal is to detrend a time series affected by non-linear trend. I do not have clear ideas about the
difference betweeen "detrending" and "smoothing". I would appreciate some suggestions and/or literature references
by people experienced about detrending and smoothing.
I was told the there are two best methods about detrending: polynomial fit on a sliding window and kernel smoothing.
I keep being confused . I got the book by Hardle. Maybe you can advice some other clarifying literature (possibly with
examples).
Thank you very much.
Maura
tutti i telefonini TIM!
[[alternative HTML version deleted]]