[R] Lags and Differences of zoo Objects
Gabor Grothendieck
ggrothendieck at gmail.com
Sat Jul 31 07:54:59 CEST 2010
On Sat, Jul 31, 2010 at 1:44 AM, pdb <philb at philbrierley.com> wrote:
>
> Thanks for the response.
>
> I can figure out the 'lag' parameter to the function, but I dont understand
> the 'differences' parameter.
>
> differences - an integer indicating the order of the difference
>
> What does the 'order of the difference' mean in English?
>
> How are these numbers calculated?
>
>> x <- iris
>> x$Species = NULL
>> x$Petal.Width = NULL
>> x$Sepal.Width = NULL
>> x$Sepal.Length = NULL
>>
>> x <- zoo(x)
>>
>> x <-
> + merge(orig = x
> + ,l1d1 = diff(x, lag = 1, differences = 1, arithmetic = TRUE, na.pad =
> TRUE)
> + ,l1d2 = diff(x, lag = 1, differences = 2, arithmetic = TRUE, na.pad =
> TRUE)
> + ,l2d1 = diff(x, lag = 2, differences = 1, arithmetic = TRUE, na.pad =
> TRUE)
> + ,l2d2 = diff(x, lag = 2, differences = 2, arithmetic = TRUE, na.pad =
> TRUE)
> + )
>>
>> x
> Petal.Length.orig Petal.Length.l1d1 Petal.Length.l1d2 Petal.Length.l2d1
> Petal.Length.l2d2
> 1 1.4 NA NA NA
> NA
> 2 1.4 0.0 NA NA
> NA
> 3 1.3 -0.1 -1.000000e-01 -0.1
> NA
> 4 1.5 0.2 3.000000e-01 0.1
> NA
> 5 1.4 -0.1 -3.000000e-01 0.1
> 2.000000e-01
> 6 1.7 0.3 4.000000e-01 0.2
> 1.000000e-01
> 7 1.4 -0.3 -6.000000e-01 0.0
> -1.000000e-01
> 8 1.5 0.1 4.000000e-01 -0.2
> -4.000000e-01
> 9 1.4 -0.1 -2.000000e-01 0.0
> 0.000000e+00
> 10 1.5 0.1 2.000000e-01 0.0
> 2.000000e-01
> 11 1.5 0.0 -1.000000e-01 0.1
> 1.000000e-01
> 12 1.6 0.1 1.000000e-01 0.1
> 1.000000e-01
> --
> View this message in context: http://r.789695.n4.nabble.com/Lags-and-Differences-of-zoo-Objects-tp2308666p2308681.html
> Sent from the R help mailing list archive at Nabble.com.
>
A kth order difference is analogous to a kth derivative and will
annihilate and (k-1)-degree polynomial. It is just repeatedly applies
diff.
z2 <- zoo((1:5)^2)
diff(z2) # linear
diff(z2, diff = 2) # constant
diff(z2, diff = 3) # all zeros
# compare
diff(z2, diff = 2)
diff(diff(z2))
More information about the R-help
mailing list