gwr_model_selection() ranks every model it evaluated on
AICc. This draws each model's criterion against its number of
predictors, the winner in red, so the gap between the best model and the
runners-up (which the ranked table shows only as numbers) is read as a
shape. A winner well below the rest is preferred; one within 2
AICc units of others is not separated from them (see "Printing" in
gwr_model_selection()). The criterion is in-sample, and the
caption says so.
Usage
# S3 method for class 'gwr_model_selection'
plot(x, ...)See also
gwr_model_selection(), which makes the object.
Other predictor selection:
gwr_model_selection(),
plot.feature_selection(),
select_features_forward()
Examples
library(sf)
set.seed(1)
n <- 80
dat <- st_as_sf(
data.frame(x = 5e5 + runif(n, 0, 1000), y = 5e6 + runif(n, 0, 1000),
a = rnorm(n), b = rnorm(n), noise = rnorm(n)),
coords = c("x", "y"), crs = 32632)
dat$z <- 2 * dat$a - dat$b + rnorm(n, 0, 0.5)
sel <- gwr_model_selection(dat, "z", c("a", "b", "noise"), bandwidth = 30)
# Every model tried, AICc against its size; the winner (a and b) marked.
plot(sel)