R/summary.R

In BAS: Bayesian Variable Selection and Model Averaging using Bayesian Adaptive Sampling

#' Print a Summary of Bayesian Model Averaging objects from BAS
#'
#' \code{summary} and \code{print} methods for Bayesian model averaging objects
#' created by \code{bas} Bayesian Adaptive Sampling
#'
#' The print methods display a view similar to \code{print.lm} .  The summary
#' methods display a view specific to Bayesian model averaging giving the top 5
#' highest probability models represented by their inclusion indicators.
#' Summaries of the models include the Bayes Factor (BF) of each model to the
#' model with the largest marginal likelihood, the posterior probability of the
#' models, R2, dim (which includes the intercept) and the log of the marginal
#' likelihood.
#'
#' @aliases print.bas print
#' @param x object of class 'bas'
#' @param digits optional number specifying the number of digits to display
#' @param ... other parameters to be passed to \code{print.default}
#' @author Merlise Clyde \email{clyde@@stat.duke.edu}
#' @seealso \code{\link{coef.bas}}
#' @keywords print regression
#' @examples
#' 
#' library(MASS)
#' data(UScrime)
#' UScrime[, -2] <- log(UScrime[, -2])
#' crime.bic <- bas.lm(y ~ ., data = UScrime, n.models = 2^15, prior = "BIC", initprobs = "eplogp")
#' print(crime.bic)
#' summary(crime.bic)
#' @rdname print.bas
#' @method print bas
#' @export
print.bas <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste(deparse(x$call),
    sep = "\n",
    collapse = "\n"
  ),
  "\n\n",
  sep = ""
  )
  cat("\n Marginal Posterior Inclusion Probabilities: \n")
  out <- x$probne0
  names(out) <- x$namesx
  print.default(format(out, digits = digits),
    print.gap = 2L,
    quote = FALSE
  )
  invisible()
}



#' Summaries of Bayesian Model Averaging objects from BAS
#'
#' \code{summary} and \code{print} methods for Bayesian model averaging objects
#' created by \code{bas} Bayesian Adaptive Sampling
#'
#' The print methods display a view similar to \code{print.lm} .  The summary
#' methods display a view specific to Bayesian model averaging giving the top 5
#' highest probability models represented by their inclusion indicators.
#' Summaries of the models include the Bayes Factor (BF) of each model to the
#' model with the largest marginal likelihood, the posterior probability of the
#' models, R2, dim (which includes the intercept) and the log of the marginal
#' likelihood.
#'
#' @aliases summary.bas summary
#' @param object object of class 'bas'
#' @param n.models optional number specifying the number of best models to
#' display in summary
#' @param ... other parameters to be passed to \code{summary.default}
#' @author Merlise Clyde \email{clyde@@duke.edu}
#' @seealso \code{\link{coef.bas}}
#' @keywords print regression
#' @examples
#' data(UScrime, package = "MASS")
#' UScrime[, -2] <- log(UScrime[, -2])
#' crime.bic <- bas.lm(y ~ ., data = UScrime, n.models = 2^15, prior = "BIC", initprobs = "eplogp")
#' print(crime.bic)
#' summary(crime.bic)
#' @rdname summary
#' @family bas methods
#' @method summary bas
#' @export

summary.bas <- function(object, n.models = 5, ...) {
  best <- order(-object$postprobs)
  n.models <- min(n.models, length(best))
  best <- best[1:n.models]
  x <- cbind(
    list2matrix.which(object, best),
    exp(object$logmarg[best] - max(object$logmarg[best])),
    round(object$postprobs[best], 4),
    round(object$R2[best], 4),
    object$size[best],
    object$logmarg[best]
  )
  x <- t(x)
  x <- cbind(NA, x)

  x[1:object$n.vars, 1] <- object$probne0
  colnames(x) <- c("P(B != 0 | Y)", paste("model", 1:n.models))
  rownames(x) <- c(object$namesx, "BF", "PostProbs", "R2", "dim", "logmarg")
  return(x)
}