archived/funcs.R

In DongyueXie/smashrgen: Empirical Bayes smoothers

#' mean squared error
#' @export
mse = function(x, y) mean((x - y)^2)

#' logit function
#' @export
logit=function(x) log(x/(1-x))

#' logistics function
#' @export
logistic=function(x) exp(x)/(1+exp(x))

#' mean integrated squared error
#' @export
mise = function(x, y, r) 10000 * mean(apply(x - rep(1, r) %o% y, 1, l2norm)/l2norm(y))

#' data.frame to ggplot data
#' @param df
#' @export
df2gg=function(df){
  n=dim(df)[1]
  p=dim(df)[2]
  name=colnames(df)
  MSE=as.vector(unlist(df))
  method=factor(rep(colnames(df),each=n),levels = name)
  dfgg=data.frame(MSE=MSE,method=method)
  return(dfgg)
}