R/biglm.R

In tidyverse/broom: Convert Statistical Objects into Tidy Tibbles

#' @templateVar class biglm
#' @template title_desc_tidy
#'
#' @param x A `biglm` object created by a call to [biglm::biglm()] or
#'   [biglm::bigglm()].
#' @template param_confint
#' @template param_exponentiate
#' @template param_unused_dots
#'
#' @evalRd return_tidy(regression = TRUE)
#'
#' @examplesIf rlang::is_installed("biglm")
#'
#' # load modeling library
#' library(biglm)
#'
#' # fit model -- linear regression
#' bfit <- biglm(mpg ~ wt + disp, mtcars)
#'
#' # summarize model fit with tidiers
#' tidy(bfit)
#' tidy(bfit, conf.int = TRUE)
#' tidy(bfit, conf.int = TRUE, conf.level = .9)
#'
#' glance(bfit)
#'
#' # fit model -- logistic regression
#' bgfit <- bigglm(am ~ mpg, mtcars, family = binomial())
#'
#' # summarize model fit with tidiers
#' tidy(bgfit)
#' tidy(bgfit, exponentiate = TRUE)
#' tidy(bgfit, conf.int = TRUE)
#' tidy(bgfit, conf.int = TRUE, conf.level = .9)
#' tidy(bgfit, conf.int = TRUE, conf.level = .9, exponentiate = TRUE)
#'
#' glance(bgfit)
#'
#' @export
#' @family biglm tidiers
#' @seealso [tidy()], [biglm::biglm()], [biglm::bigglm()]
tidy.biglm <- function(x, conf.int = FALSE, conf.level = .95,
                       exponentiate = FALSE, ...) {
  # TODO: separate in biglm and bigglm tidiers

  ret <- as_tibble(summary(x)$mat, rownames = "term")
  colnames(ret) <- c("term", "estimate", "conf.low", "conf.high", "std.error", "p.value")

  # remove the 95% confidence interval and replace:
  # it isn't exactly 95% (uses 2 rather than 1.96), and doesn't allow
  # specification of confidence level in any case
  ret <- dplyr::select(ret, -conf.low, -conf.high)

  if (conf.int) {
    ci <- broom_confint_terms(x, level = conf.level)
    ret <- dplyr::left_join(ret, ci, by = "term")
  }

  if (exponentiate) {
    ret <- exponentiate(ret)
  }

  ret
}


#' @templateVar class biglm
#' @template title_desc_glance
#'
#' @inherit tidy.biglm params examples
#' @template param_unused_dots
#'
#' @evalRd return_glance("r.squared",
#'                       "AIC",
#'                       "deviance",
#'                       "df.residual",
#'                       "nobs")
#'
#' @export
#' @family biglm tidiers
#' @seealso [glance()], [biglm::biglm()], [biglm::bigglm()]
glance.biglm <- function(x, ...) {
  s <- summary(x)
  as_glance_tibble(
    r.squared = s$rsq,
    AIC = stats::AIC(x),
    deviance = stats::deviance(x),
    df.residual = x$df.resid,
    nobs = stats::nobs(x),
    na_types = "rrrii"
  )
}