Displaying 2 results from an estimated 2 matches for "dcomp".
Did you mean:
comp
2011 Dec 13
1
Rcpp too good to be true?
...im pretty new to
this) it would be much appreciated. I've had a simple look at the results
and they look fine, but seriously, 500x faster?!
function in R;
library(compiler)
Rrcomp <- cmpfun(
function(n, lam, nu, max = 100L) {
ans <- integer(n)
dist <- dcomp(0:max, lam, nu, max)
u <- runif(n)
for (i in 1:n) {
count <- 0L
pr <- dist[1L]
while (pr < u[i]) {
count <- count + 1L
pr <- pr + dist[count + 1L]
}...
2005 Nov 27
1
fixed, random effects with variable weights
...ld give me a hint, this would be
really great.
Basically, I want to run a standard economist's fixed, and random
effects regression (corresponds to xtreg in STATA) but with _variable_
weights (they correspond to changing industry shares in the market).
Here is what I do:
regsc<-lme(dsc~dcomp+dperc,random=~1|ind7090)
update(regsc,weights=varFixed(~wt))
1. however, my results are different from what I obtain in Stata using
areg (the weighted fixed effects times series regression). any ideas?
2. how do I read of the random affects results from this regression?
(i.e. coefficients on dco...