Autoregressive AR models are used a lot in forecasting and prediction.
The problem with standard AR models is that they use normal Gaussian errors.
We all know that financial returns are not normally distributed.
We can use Bayesian analysis and build AR models that have long tail returns.
We use student t distribution to model the long tail behavior of the financial returns.
I have written a blog post in which I explain in detail how we do it using R and BUGS.
BUGS is a powerful Bayesian Analysis software developed by Cambridge University.
BUGS revolutionized the field of Bayesian Statistics.
Now we can use BUGS in building trading models by connecting it with R.
Read the blog post in which I explain in detail how to do it.