R/fns.R

In mas3321: Bayes normal linear model and basic MCMC

#' Function calculates a quadratic form for 2-d vectors
#' @param beta1 dimension 1
#' @param beta2 dimension 2
#' @param b 2-d vector containing (location) mean
#' @param c 2x2 centering matrix
#' @export quadform2d
#' @examples quadform2d(1,1,c(0,0),diag(1,2))
quadform2d=function(beta1,beta2,b,c){
  c[1,1]*(beta1-b[1])^2+c[2,2]*(beta2-b[2])^2+2*c[1,2]*(beta1-b[1])*(beta2-b[2])}

def.par <- par(no.readonly = TRUE) # save default, for resetting...

#' Parameters of the NGa posterior distribution when sampling from a normal population
#' and using a NGa prior distribution
#' @param priorpara vector of prior parameters b,c,g,h
#' @param n sample size
#' @param xbar sample mean
#' @param s sample standard deviation
#' @export NGaposterior
#' @examples NGaposterior(c(0,1,1,1),10,1,1)
NGaposterior=function(priorpara,n,xbar,s){b=priorpara[1];c=priorpara[2];g=priorpara[3];h=priorpara[4];c((b*c+n*xbar)/(c+n),c+n,g+n/2,h+c*n*(xbar-b)^2/(2*(c+n))+n*s^2/2)}

#' Launches an editor containing the commands in a demo
#' @param demo name of demo
#' @export demoCommands
demoCommands=function(demo){
  edit(file=system.file("demo", paste(deparse(substitute(demo)),".R",sep=""), package="mas3321"))
}