Skip to content

Code for Bayesian inference for queueing networks with incomplete data

License

Notifications You must be signed in to change notification settings

casutton/bayes-qnet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

bayes-qnet

Code for Bayesian inference for queueing networks with incomplete data.

The statistical methods are described in this paper:

C. Sutton and M. I. Jordan. Bayesian inference in queueing networks. Annals of Applied Statistics, 5(1):254–282, 2011. https://arxiv.org/abs/1001.3355

System Requirements

This code requires Python 2 and scipy. Also requires yaml.py

Significant portions of the time critical portions of the code are written in Cython. The C translations of the Cython modules are included in the git repo.

Installation

You will need to create the C shared libraries to use the Python code. If you have Cython properly installed, this can be done using

  cd src/
  make

Getting Started

Of course, there is no real documentation, but I have created a Jupyter notebook that walks through an example usage. This is in the examples/ directory.

I've saved the notebook both in native Jupyter format, so you can run the code one your machine, and as a static HTML page, in case you do not have Jupyter installed.

Hopefully this is enough to get started.

Test Suite

(unfortunately this part is from memory)

To test if you have compiled correctly, you can run the unit tests. These are all in the src/ directory and their names begin with test_. Why I decided not to put them in separate directory, who knows.

The most important one is test_qnet.py. So just

  cd src/
  python2 test_qnet.py

should work. A dozen of the tests will fail because I just removed some unused (but moderately testing) functionality to make the package a lot easier to install.

About

Code for Bayesian inference for queueing networks with incomplete data

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published