luisdamiano / bayeshmm Goto Github PK

First of all: great package!

I have been looking at how the BayesHMM package would support panel data but so far I am stuck. I was wondering if you could help me out.

Recall: contrary to a time-series dataset with just a time-dimension, panel data includes both the time-dimension and an observational dimension. A typical example of panel data is following N customers over T time periods.

Thank you in advance.

Hi!
Really love the pacakge, looks awesome!
Im on an arch linux machine.

im getting an error -

Thank you @luisdamiano

Best,
Andrew

Decide if the content in R/write.R and R/writer.R is duplicated and consolidate into one file.

I needed to update rstan.

I get the following error when running the example hmm-univariate-gaussian.R:

> myOpt   <- optimizing(mySpec, myModel, y = y, nRun = 50, nCores = 10, seed = 9000)
Error in (function (file, model_name = "anon_model", model_code = "",  : 
  unused argument (seed = 9000)

Removing seed = 9000 worked. Is seed still applicable in this function?

Hi,
I run example"hmm-multivariate-gaussian-priors.R", get "Error: passing unknown arguments: stanModel" for cmds:
myModel <- compile(mySpec)
myFit <- draw_samples(mySpec, stanModel = myModel, y = y, chains = 1, iter = 500, seed = 9000)

the first line pass without error.
but the second failed.

What would you suggest?

BR
ZSY

Every time when I used the sampling function, I got this issue. And I try find the reason and solutions, there hava a comment says is the problem of storage, the original function saves many data each sampling, so the R session crashes, and i try small iter number then there no problem, or set par="" of sampling funciton, means save no parameters. It's that right? Which way should I apply to solve this problems? Thanks.