R/performance_mse.R

In performance: Assessment of Regression Models Performance

#' @title Mean Square Error of Linear Models
#' @name performance_mse
#'
#' @description Compute mean square error of linear models.
#'
#' @inheritParams performance_rmse
#' @inheritParams model_performance.lm
#'
#' @details
#' The mean square error is the mean of the sum of squared residuals, i.e. it
#' measures the average of the squares of the errors. Less technically speaking,
#' the mean square error can be considered as the variance of the residuals,
#' i.e. the variation in the outcome the model doesn't explain. Lower values
#' (closer to zero) indicate better fit.
#'
#' @return Numeric, the mean square error of `model`.
#'
#' @examples
#' data(mtcars)
#' m <- lm(mpg ~ hp + gear, data = mtcars)
#' performance_mse(m)
#' @export
performance_mse <- function(model, ...) {
  UseMethod("performance_mse")
}

#' @rdname performance_mse
#' @export
mse <- performance_mse



#' @export
performance_mse.default <- function(model, verbose = TRUE, ...) {
  res <- .safe(insight::get_residuals(model, verbose = verbose, type = "response", ...))

  if (is.null(res)) {
    res <- .safe({
      def_res <- insight::get_residuals(model, verbose = FALSE, ...)
      if (verbose) {
        insight::format_alert(
          "Response residuals not available to calculate mean square error. (R)MSE is probably not reliable."
        )
      }
      def_res
    })
  }

  if (is.null(res) || all(is.na(res))) {
    return(NA)
  }

  # for multivariate response models...
  if (is.data.frame(res)) {
    if (verbose) {
      insight::format_warning("Multiple response variables detected. Cannot reliably compute (R)MSE.")
    }
    return(NA)
  }

  mean(res^2, na.rm = TRUE)
}




# mfx models -------------------------------

#' @export
performance_mse.logitor <- function(model, verbose = TRUE, ...) {
  performance_mse(model$fit, verbose = verbose, ...)
}

#' @export
performance_mse.logitmfx <- performance_mse.logitor

#' @export
performance_mse.probitmfx <- performance_mse.logitor

#' @export
performance_mse.poissonirr <- performance_mse.logitor

#' @export
performance_mse.poissonmfx <- performance_mse.logitor

#' @export
performance_mse.negbinirr <- performance_mse.logitor

#' @export
performance_mse.negbinmfx <- performance_mse.logitor

#' @export
performance_mse.betaor <- performance_mse.logitor

#' @export
performance_mse.betamfx <- performance_mse.logitor

#' @export
performance_mse.model_fit <- performance_mse.logitor