Why does rjags stop when trying to fit y to gamma distribution?
That approach isn’t working for our data because when rjags tries to fit a negative value of y to a gamma distribution, it stops. In concept perhaps one could route all negative observations to the normal distribution centered near zero and let rjags distribute the rest. It would involve a little fancy footwork with the distribution results.
What are the aims of the Jags language?
JAGS was written with three aims in mind: to have an engine for the BUGS language that runs on Unix; to be extensible, allowing users to write their own functions, distributions, and samplers; and to be a platform for experimentation with ideas in Bayesian modelling.
When to use Jags returns-INF when calculating log density?
JAGS returns -Inf when calculating the log density associated with negative values from the Gamma distribution, which is as expected (i.e. the log of a probability of 0), so I’m not sure exactly what error you got when you tried it.
How to do a Gaussian-gamma mixture in MATLAB?
There’s an unanswered question about how to do this in Matlab at EM of a Gaussian-Gamma mixture model in matlab. Package bmixture provides for Bayesian mixture models of multiple gammas or multiple normal distributions but not a combination of both distributions.
When to use mixture model in statistical modelling?
Often in modelling of statistical processes with basic distributions, like Gaussian, Beta or Gamma, the observed data is not captured very well by the chosen distribution. What is asked for, is often a mixture of distributions that, taken together, fit the observed data more closely.
How does a two component mixture model work?
A two-component mixture model of different parametric distributions. The vector probs will store the probabilities alpha and 1-alpha. Its prior distribution follows a dirichlet distribution of order two. A normal prior is put on the scalars lower, upper and mu, and a gamma prior is set for tau.