Hi friend

I have worked on the feature selection methods in my papers including rough set theory.You may see it.

Best regards

Dr.Indrajit Mandal

Hi,

I think you are mixing up things since Bayesian networks, feature selection and markov blanket are interrelated but very things. I guess you are asking about using markov blanket to do feauture selection in a Bayesian network. I dont know about any implementation in Matlab, but if you read about the concept, you will see that any sofware that allows to search for the graph of a BN can be used:

http://en.wikipedia.org/wiki/Markov_blanket

I would recommend you also the following article about different feature selection methods:

http://bioinformatics.oxfordjournals.org/content/23/19/2507.short

Hope this helps

I believe the link below will help

http://link.springer.com/chapter/10.1007%2F978-3-642-22910-7_7

Hi Tahseen, did you succeed in your searches? Just came up with the same quastion.

Hello Asatur,

The Markov Blanket is like causality testing. So as far as the series are equal length a number of pattern based or distance based techniques are provided in hClust, tsClust and dtwclust packages in R. As, i am applying these in life science and did in meta- genomics, so didn't try the rough sets.

P-value based Markov Blanket is also possible by changing the model based comparison in dtwClust and modified with auto.arima in tsClust to accommodate non-linear models.

https://www.rdocumentation.org/packages/fastcluster/versions/1.1.25/topics/hclust

https://cran.r-project.org/web/packages/TSclust/TSclust.pdf

https://cran.r-project.org/web/packages/dtwclust/dtwclust.pdf

Thank you for your answer. I was searching for a matlab implementation as it is mentioned in your question. Thank you.