|
Has there been a comprehensive (with a non-toy dataset such as UCI) implementation of Belief network using C#? I am interested in an implementation which handles building a deep belief network possibly with persistence capabilities. Thanks in advance. |
|
Have you checked out Infer.net? It build all sorts of graphical models (including bayesian and Markov networks) and performs inference in many different ways. It's quite efficient as it compiles your model down into CLI bytecode, which can then be JIT'd. @Adnam Mascod: the standard distribution of Infer.NET comes with many examples built-in, and all these examples have code to show you how to set the observed variables in the model and run inference on them. While they don't load from any public datasets (probably due to licensing issues), as long as you know how to open a file in C# and read strings from it you should be fine when populating such a model yourself. If you still can't manage to get something running, post example code and data at the infer.NET forum and people over there will certainly help you out.
(May 11 '11 at 16:27)
Alexandre Passos ♦
|
|
Thanks for the response Alexandre. I have done some basic experiments and prototyping with infer.net. However, I couldn't find a good example of it loading and building a graphical model using a public dataset. It's not available on MSR but if you know any such resource, that would be great. |
|
I know this answer is very late, but if your still interested you might want to check out Accord.Net. There is a overview, youtube intro video and application of the Deep Belief network in c# available here: Deep Neural Networks and Restricted Boltzmann Machines. |
|
Hello all, I want to implementat Restricted boltzmann Machine in C# with example dataset : http://www.ieor.berkeley.edu/~goldberg/jester-data/ please help me! My email: [email protected] |
