Testing truncated Normal sampling
Truncated Normal latents with non-unit variance are necessary for logistic BART.
draw_lambda_i(lambda, mean, kmax=1000, thin=1)
lambda
: Previous value of lambda.mean
: Mean of truncated Normal.kmax
: The number of terms in the mixture.thin
: The thinning parameter.Returns the variance for a truncated Normal, i.e., c("", "").
rtnorm, lbart
set.seed(12) draw_lambda_i(1, 2) rtnorm(1, 2, sqrt(6.773462), 6) draw_lambda_i(6.773462, 2)
Useful links