R/approxBF.R

In masspastore/ADati: Analisi dei dati in Psicologia

#' Bayes Factor approssimato
#' @param M0 = Modello nullo
#' @param M1 = Modello target
approxBF <- function(M0,M1) {
  (Bic0 <- BIC(M0))
  (Bic1 <- BIC(M1))
  (dBic <- Bic0-Bic1)
  (BF <- exp(dBic/2))
  
  #(support <- ifelse(dBic<=0,"supports M0","supports M1"))
  support <- "supports model"

  cat("-------------------","\n")
  cat(paste("Bayes Factor",support),"\n")
  if (dBic<=0) print(formula(M0))
  if (dBic>0) print(formula(M1))
  
  cat("-------------------","\n")
  cat(" ","\n")
  return(list(dBIC=dBic,BF=BF))
}

#'@examples
#'x <- sample(0:1,10,TRUE)
#'y <- rnorm(10)
#'M0 <- lm(y~1)
#'M1 <- lm(y~x)
#'approxBF(M0,M1)
#'approxBF(M1,M0)