So many chains for too little iterations.
Check Bob suggestion too.
2014/1/18 Bob O'Hara boh...@senckenberg.de
On 01/18/2014 04:39 AM, æ wrote:
Dear all
I have a model with four paramters, I want to estimate the
parameter uncertainty, so Bayesian analysis with MCMC method is
David,
I was stuck with a similar problem. I tried creating LUCC models and R
can't do the job in the required speed and or just stop working because of
lack of memory.
Problem with R is default type is double even for integers and this
behaviour will eat all memory very fast for 10ⶠagents,
*error for node smtree[1] of type GraphPoisson.Node first argument must be
positive*
You're using negative (must be positive) value to lambda parameter in
Poisson distribution.
You should start checking this line:
*mean[i] - alpha[site[i]] + re[i]
*
and find why mean[i] is returning negative
You can check reshape or melt functions. They do exactly what you want.
2012/6/29 Baldwin, Jim -FS jbald...@fs.fed.us
Sorry. It looked fine when I sent it. Below I've added back in the line
feeds that somehow disappeared.
Jim
-Original Message-
From:
Maybe your solution can be found here:
http://cran.r-project.org/doc/FAQ/R-FAQ.html#Why-does-the-output-from-anova_0028_0029-depend-on-the-order-of-factors-in-the-model_003f
2012/4/3 Kristen Gorman kgor...@sfu.ca
Dear all,
I have R code to run AIC including multi-model inference. I am running
Algorithms for bayesian inference (MCMC) cannot run fast enough inside a
scripting language like R.
Most authors create plugins to call their fast binary (C/C++)
implementations inside of R. Just use them.
I recommend JAGS and rjags. Better errors messages and good support from
Plummer.
Cya
Hi Mahnaz,
You didn't state your problem.
How to ask smart questions: http://catb.org/esr/faqs/smart-questions.html
2012/1/31 Mahnaz Rabbaniha rab.mah...@gmail.com
Dear r-project member
hi
i have problem to use nmds in vegan in my data that is include
environmental factors (
Default packages of any installation of R can perform many types of
regression (IMO regression can show you something). Your goal seems to wide
for me, maybe you can provide more information like hypotheses?
Any google query using community and R will show *vegan* (popular package
for community
If you known a bit of programming you can change dnorm in that code to do a
proper non-normal regression. Rank method will work just if you have a
large enough sample (central limit theorem).
https://github.com/dvdscripter/R-snippets/blob/master/regression.R
2012/1/8 Mahnaz Rabbaniha
