Displaying 1 result from an estimated 1 matches for "slice1d".
Did you mean:
slice
2007 Feb 09
0
slice sampler
...I basicly followed neal 03 to write the code. The code might looks a little strange as I allowed sampling a vector of parameters , that is , it could update multiple targets with different parameter but same form of distribution. A example
of usage is as follows:
output[i] <- slice1d(dnorm(x,1,1,log=TRUE), x0=output[i-1],w=2,m=6,trunc=c(-Inf,Inf),prec=1e-4)
things to notice here are: expression has to be on LOG scale, w is the expanding wideth, m is the # of expanding trials,
prec is the precision, this is used to control against those distributions with huge kurtosis (with v...