R/Y_generator.R

In CorReg: Linear Regression Based on Linear Structure Between Variables

# ' Response variable generator with a linear model
# ' @param X the dataset
# ' @param Amax maximum number of non-zero coefficients
# ' @param sigma_Y the standard deviation of the noise
# ' @param positive the ratio of positive coefficients
# ' @param tp1 ratio of the right-side covariates in A
# ' @param tp2 ratio of the left-side covariates in A
# ' @param tp3 ratio of the totally independent covariates (if exists) in A
# ' @param pb kind of problem: 0=none, 1=simple, 2=strong
# ' @param Z the structure adjacency matrix
# ' @export 
Y_generator <- function(X = X, Amax = NULL, sigma_Y = 10, positive = 0.6, Z = NULL, tp1 = 1, tp2 = 1, tp3 = 1, pb = 0){
  p = ncol(X)
  taille = nrow(X)
#   A = rpois(p+1, 5)*(rep(-1, p+1)+2*rbinom(p+1, 1, positive))
  if(is.null(Z)){
     Z = matrix(0, ncol = ncol(X), nrow = nrow(X))
  }
  A = generateurA_ou(Z = Z, tp1 = tp1, tp2 = tp2, tp3 = tp3, positive = positive, pb = pb, Amax = Amax)
#   if(!is.null(Amax)){
#     Amax = min(abs(Amax), p)
#     A[-sample(2:(p+1), size = Amax)] = 0 # on garde la constante
#   }
  Y = cbind(rep(1, times = taille), as.matrix(X))%*%A + rnorm(taille, mean = 0, sd = sigma_Y)
  
  return(list(Y = Y, A = A))
}