[R-SIG-Finance] Estimating credit rating transition matrices

[Milos Cipovic]

Hello guys,

Has anybody done this with one of the markov chain packages? I've been
trying to find a function for hazard rate (or duration, or homogeneous
time) approach but I kinda get confused with every vignete i find. I've
tried markovchain, semimarkovchain, msm, hmm packages, and I think that
hazard rate is nested in one of models from these packages, but I'm not
shore.

I noticed that some of them use two columns for states (start state and end
state), although this is not a big deal, it would be great if I could fit
data with one column of states.

Also, in discrete case, how did you solve non rating state?

Thank you.

	[[alternative HTML version deleted]]