Code for a Bayesian GARCH(1,1) volaitility model

Kondwani · February 27th, 2013, 7:03 am

I am doing research on single regime and regime switching Bayesian GARCH volatility models. The benchmark for my research is work by David Ardia and Teruo Nakatsuma who have managed to implement these models well. My problem is coding the MCMC algorithms to estimate the parameters, using R. I need a starting point for the coding and as such I would appreciate assistance with code (preferably in R) or insight from anyone working in the same area. If code is available, it can surely help me a great deal in my research as I will build on that code and write my own version.

Cuchulainn · February 27th, 2013, 9:20 am

What about BUGS

Kondwani · February 27th, 2013, 9:29 am

BUGS is not versatile enough. My goal is to construct and fine-tune an original MCMC algorithm from scratch.

4rcher · February 27th, 2013, 9:38 am

check this package and the related paper

Cuchulainn · February 27th, 2013, 11:39 am

QuoteOriginally posted by: KondwaniBUGS is not versatile enough. My goal is to construct and fine-tune an original MCMC algorithm from scratch.Quote I need a starting point for the coding and as such I would appreciate assistance with code (preferably in R) or insight from anyone working in the same areaI see a contradiction here. Starting from scratch implies that you should not need coding help. It could be too ambitious, but I don't have all the details.How many months has been allocated to this project?

Kondwani · February 27th, 2013, 3:49 pm

I have four months to complete this task. That is I have until June 2013.