[R] Smoothing by group - Panel data - exponential/loess

[Miluji Sb]

Dear all,

I have panel data for a series (g) for three time periods. The variable is
likely autocorrelated. I would like to generate a new variable using
exponential/loess smoothing by group (gid).

For time series, I could have done something like this;

smoothdf <- data.frame(
  x = 1:n,
  y = as.vector(smooth(dat$g)),
  method = "smooth()"
)

But confused about the panel data setting. Apologies if this a stat
question along with an R query. Any help will be greatly appreciated.

### data
structure(list(gid = c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 4L,
4L, 4L, 5L, 5L, 5L, 6L, 6L, 6L, 7L, 7L, 7L, 8L, 8L, 8L, 9L, 9L,
9L), year = c(2030L, 2050L, 2100L, 2030L, 2050L, 2100L, 2030L,
2050L, 2100L, 2030L, 2050L, 2100L, 2030L, 2050L, 2100L, 2030L,
2050L, 2100L, 2030L, 2050L, 2100L, 2030L, 2050L, 2100L, 2030L,
2050L, 2100L), g = c(9.24e-05, 0.0001133, 6.3e-05, 7.72e-10,
1.41e-09, 4.68e-09, 0.0001736, 0.0002286, 0.0001541, 1.87e-15,
3.76e-15, 8.52e-15, 0.0001822, 0.0002391, 0.0001348, 3.69e-06,
8.11e-06, 8.63e-06, 3.41e-06, 7.32e-06, 7.18e-06, 8.47e-07, 1.83e-06,
1.84e-06, 1.13e-06, 2.37e-06, 2.15e-06)), class = "data.frame", row.names =
c(NA,
-27L))


Sincerely,

Milu

	[[alternative HTML version deleted]]