|
|
| Author |
Message |
comtech science forum Guru Wannabe
Joined: 08 Oct 2005
Posts: 150
|
Posted: Mon May 22, 2006 10:07 pm Post subject:
Does strong Markov property impley Markov property?
|
|
|
Dear all,
I understand that Markov property does not imply strong markov
property; but strong markov property implies markov property.
Is my understanding correct?
Thanks a lot! |
|
| Back to top |
|
 |
BuddhaThu science forum beginner
Joined: 06 Oct 2005
Posts: 21
|
Posted: Wed May 24, 2006 7:30 pm Post subject:
Re: Does strong Markov property impley Markov property?
|
|
|
Dear Comtech,
I am not an expert in Markov, but let me take a crack at this.
In interpreting time, you can see it as either discrete or continuous.
This might be the same for interpreting Markov.
There are two types of Markov.
The first type is discrete Markov where you can definitively demarcate
between the past, present and future. It is sort like Monday, Tuesday
and Wednesday with Monday not considered as having causal relation to
Tuesday and Wednesday.
Note: The past is not deleted. It is just not considered into the
relationship.
This might be considered as strong Markov.
The other type is a continuous Markov chain. Here, since it is
continuous, you can only approximate the time difference between the
causal/statistical chain of present to future, with the approximate
past sliced off and not considered.
Note: There is no definitively or objective way to cut the past. There
is only direction and magnitude using vectors.
It is like the Dedekind cut where the cut is created by the
mathematician.
In my interpretation, since it is 'continuous,' there is no real
'cut' at all from the past, even if it is by the **subjective**
hand of the mathematician. This is very different from the discrete
Markov chain.
With the continuous chain, you can only approximate it as the immediate
past and arbitrarily mark it off not to be considered. The past will
always there no matter how minute you cut.
No one can 'cut' an infinite continuum. There is no precise
'cut.'
Hence, the Markov property does not imply strong Markov, while strong
Markov will imply the Markov property.
The strong is a discrete form of the Markov property so that there can
be a 'precision' discrete cut.
The property alone only has the continuous and can only approximate the
cut from the past.
B.T. |
|
| Back to top |
|
 |
C6L1V@shaw.ca science forum Guru
Joined: 23 May 2005
Posts: 628
|
Posted: Wed May 24, 2006 8:50 pm Post subject:
Re: Does strong Markov property impley Markov property?
|
|
|
comtech wrote:
| Quote: | Dear all,
I understand that Markov property does not imply strong markov
property; but strong markov property implies markov property.
Is my understanding correct?
|
Basically, YES. However, there is more: for the case of a homogeneous,
finite-state continuous time Markov chain, one can replace the process
by an equivalent strong Markov process; this is a theorem that is
proved in various advanced treatments. (Equivalent means that all the
finite-dimensional joint distributions are the same.) Therefore, one
might as well assume the strong Markov property right from the start. I
don't know if this result still holds for countably-infinite state
chains in continuous time, or even in discrete time.
By the way, response #2 in this thread is pure nonsense. The strong
Markov property is not concerned much with whether or not time is
continuous.
R.G. Vickson
|
|
| Back to top |
|
 |
Google
|
|
| Back to top |
|
 |
|
|
The time now is Thu Mar 11, 2010 5:56 pm | All times are GMT
|
|
Copyright © 2004-2005 DeniX Solutions SRL
|
|
Other DeniX Solutions sites:
Electronics forum |
Medicine forum |
Unix/Linux blog |
Unix/Linux documentation |
Unix/Linux forums
|
Powered by phpBB © 2001, 2005 phpBB Group
|
|