Bayesian networks tutorial [closed]

StackOverflow https://stackoverflow.com/questions/345280

Question

For a beginner, which is the best book to start with for studying Bayesian Networks?

Was it helpful?

Solution

I would recommend "Probabilistic Graphical Models" by Daphne Koller and Nir Friedman. Its an excellent starter-to-intermediate handbook on both directed (Bayesian Networks) and undirected (Markov Networks) graphical models. The examples given are elaborate and easy to understand.

OTHER TIPS

A good book on general machine learning is 1. But it is quite light on BN. I haven't read [2] but I have read [3] by him which is good (so, [2] is likely to be good as recommended by dwf). I would not recommend Pearl's book at all unless you are doing your Ph.D.!

However, I actually would recommend the online tutorial "A Brief Introduction to Graphical Models and Bayesian Networks" by Kevin Murphy [4]. The best way to learn BN is to read this, download his Matlab toolbox [5] and build your own BN in ten minutes.

  1. Pattern classification by Duda/Hart/Stork
  2. Pattern Recognition and Machine Learning by Chris Bishop
  3. Neural Networks for Pattern Recognition by Chris Bishop
  4. http://www.cs.ubc.ca/~murphyk/Bayes/bnintro.html
  5. Bayes Net Toolbox for Matlab

You should check for AI (Artificial Intelligence) books. I've learn about Bayesian in Artificial Intelligence "A modern approach".

This online book has been extremely helpful for me in all aspects of machine learning, including Bayesian inference:

http://www.inference.phy.cam.ac.uk/mackay/itila/book.html

Granted you are familiar with basic probability theory, its a great resource.

All the books mentioned so far are pretty good ones. Pearl's is generally regarded as being a bit hard to follow, it's also quite expensive, but if you can manage it, all the power to you.

I'd really really recommend you check out Chris Bishop's book, Pattern Recognition and Machine Learning. I think it's far and away the best treatment you're going to get of graphical models in a textbook, at least until Michael Jordan finishes and publishes his book on the subject.

The best professors in this fields are by my point of view these 2 guys:link text Ng. Andrew and link text Prof. Pallab Dasgupta.

I have been watching all their tutorials on BBN and they were very usefull.Just follow the links and you will find more AI lectures with this 2 interesting guys.

Have fun learning with them, Mike

Pearl's 1988 Probabilistic Reasoning in Intelligent Systems is the one of the most cited works on Bayesian Networks. I found it quite clear. That said, a lot has been done in the field since 1988. It would be wise to supplement this book with more recent works.

Mitchell's Machine Learning is an extremely important primer in the area of AI. It covers Bayesian Networks, devoting, as I recall, an entire chapter to it.

I'd also check out Weka's Bayesian Network class to understand a practical implementation. If you don't know about Weka, check it out here: http://www.cs.waikato.ac.nz/ml/weka/

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top