R/log_likelihood.R

In CorReg: Linear Regression Based on Linear Structure Between Variables

# ' logvraisemblance d'une sous-r?gression
# '
# ' @param theta vector of the parameters c(sigma,beta)
log_likelihood <- function(theta = theta, Y = Y, X = X) {
  X = as.matrix(X) # cbind(rep(1,times=nrow(X)),X)
  return(-(length(Y) / 2) * log(2 * pi * theta[1]^2) - (sum((Y - as.matrix(X) %*% theta[-1])^2)) / (2 * theta[1]^2))
}