local_slope: Calculate the local slope of a smooth term using finite...

In bentrueman/bgamcar1: Fit Bayesian GAMs With CAR(1) Errors To Censored Data

Calculate the local slope of a smooth term using finite differences

Description

Calculate the local slope of a smooth term using finite differences

Usage

local_slope(
  input,
  object,
  x_var,
  epsilon = 0.001,
  smooth,
  g_var = NULL,
  add_vars = NULL,
  pts = 200,
  ...
)

Arguments

input	A dataframe for which to generate model predictions.
object	A brms model object.
x_var	The predictor associated with the selected smooth term.
epsilon	The increment on which to generate differences.
smooth	The smooth term for which to calculate local slopes. Passed on to brms::posterior_smooths().
g_var	An optional grouping variable for factor-smooth interactions.
add_vars	In older versions of brms, brms::posterior_smooths() required that all variables used in constructing the autocorrelation term are also included in newdata arg... even if they are not used in the smooth. Add them here as a named list of new columns; if they don't appear in the smooth, they won't contribute to the output.
pts	Number of points at which to estimate the slope.
...	Additional arguments passed to brms::posterior_smooths().

Value

A tibble containing the (tidy) output of brms::posterior_smooth() and the calculated slopes.

Examples

library("brms")
seed <- 1
data_gam <- read.csv(paste0(system.file("extdata", package = "bgamcar1"), "/data_gam.csv"))
fit <- fit_stan_model(
  paste0(system.file("extdata", package = "bgamcar1"), "/test_gam"),
  seed,
  bf(y ~ s(x0) + s(x1) + s(x2) + s(x3)),
  data_gam,
  car1 = FALSE,
  chains = 2
)
local_slope(data_gam, fit, "x2", smooth = "s(x2)")

bentrueman/bgamcar1 documentation built on July 6, 2024, 11:16 p.m.