Histograms of observed and predicted response — plot_prediction

Compares the distribution of the observed response with the distribution of predicted values over the prediction grid, on a log1p x-axis, with the means marked. Blue lines indicate mean values.

If the model has an offset, the observed response is divided by exp(offset) (e.g. catch per unit swept area), so it is on the same scale as predictions made without an offset (as for a prediction grid).

Usage

plot_prediction_histogram(
  fit,
  pred,
  response = NULL,
  column = "est",
  base_size = 11
)

Arguments

fit: A fitted sdmTMB::sdmTMB() model (for the observed response).
pred: A prediction data frame from predict(fit, newdata = ...).
response: Character. Name of the response column in the model data. If NULL (default), taken from the model formula.
column: Character. Predicted column in pred. Default "est".
base_size: Numeric. Base font size. Default 11.

Value

A cowplot grid of two ggplot panels.

Author

Mikko Vihtakari

Examples

# \donttest{
mesh <- sdmTMB::make_mesh(sdmTMB::pcod, c("X", "Y"), cutoff = 20)
m <- sdmTMB::sdmTMB(
  density ~ 0 + as.factor(year), data = sdmTMB::pcod, time = "year",
  mesh = mesh, family = sdmTMB::tweedie()
)
nd <- sdmTMB::replicate_df(sdmTMB::qcs_grid, "year", unique(sdmTMB::pcod$year))
pred <- predict(m, newdata = nd, type = "response")
plot_prediction_histogram(m, pred)

# }