Jacknife resampling
# S4 method for FLQuant jackknife(object, dim = "year", rm.na = TRUE)
The jacknife
method sets up objects ready for jacknifing, i.e. to
systematically recompute a given statistic leaving out one observation at a
time. From this new set of "observations" for the statistic, estimates for
the bias and variance of the statstic can be calculated.
Input objects cannot have length > 1 along the iter
dimension, and
the resulting object will have one more iter
than the number of
elements in the original object.
jacknife(object, ...)
#> -- iter: 1 #> 1 2 3 4 5 6 7 8 #> 1 2 3 4 5 6 7 8 #> -- iter: 2 #> 1 2 3 4 5 6 7 8 #> NA 2 3 4 5 6 7 8 #> -- iter: 3 #> 1 2 3 4 5 6 7 8 #> 1 NA 3 4 5 6 7 8 #> -- iter: 4 #> 1 2 3 4 5 6 7 8 #> 1 2 NA 4 5 6 7 8 #> -- iter: 5 #> 1 2 3 4 5 6 7 8 #> 1 2 3 NA 5 6 7 8 #> -- iter: 6 #> 1 2 3 4 5 6 7 8 #> 1 2 3 4 NA 6 7 8 #> -- iter: 7 #> 1 2 3 4 5 6 7 8 #> 1 2 3 4 5 NA 7 8 #> -- iter: 8 #> 1 2 3 4 5 6 7 8 #> 1 2 3 4 5 6 NA 8 #> -- iter: 9 #> 1 2 3 4 5 6 7 8 #> 1 2 3 4 5 6 7 NA #> #> units: NA#> NULL# Calculate the bias of the mean and variance estimators (mean(iter(yearMeans(flj),2:9))-c(iter(yearMeans(flj),1)))*7#> [1] 0#> [1] 0