Skip to contents

Estimate NSEC values from a multivariate brmsfit

Source: R/nsec_multi.R
nsec_multi.Rd

[Defunct]

nsec_multi() is defunct and calling it is an error. Use nsec(), which gives one row per response and direction for a multivariate fit, reading the responses from the fit.

Usage

nsec_multi(...)

Arguments

...

Unused.

Converting to nsec()

multi_var and trials_var are not needed: the responses and any trials column are read from the fit. type becomes a filter on the direction column, and posterior = TRUE becomes draws = TRUE.

# nsec_multi(fit, x_var = "dose", multi_var = "sp")
nsec(fit, x_var = "dose")

# nsec_multi(fit, x_var = "dose", multi_var = "sp", type = "decreasing")
nsec(fit, x_var = "dose") |>
  dplyr::filter(direction == "decreasing")

# nsec_multi(fit, x_var = "dose", multi_var = "sp", type = "lower",
#   criterion = 0.8)
nsec(fit, x_var = "dose", draws = TRUE) |>
  dplyr::mutate(
    at_criterion = vapply(
      draws,
      \(d) quantile(d$value.draw, 0.8, na.rm = TRUE),
      numeric(1)
    )
  ) |>
  dplyr::filter(!is.na(estimate)) |>
  dplyr::group_by(response) |>
  dplyr::slice_min(at_criterion, n = 1, with_ties = FALSE) |>
  dplyr::ungroup()