This package depends on the ThinkBayes code package. We already include the parts of the package (as of 2015) that we use in this repository in the thinkbayes.py, but you may wish to update it to the latest version of it from ThinkBayes' Github repository.

Once you have this installed, you can either try to run your own tests using the instructions below, or run the Benchmarks

For purposes of testing, we encapsulate norm detection algorithms in the norm_detector class, which is then subclassed to implement different norm detection strategies, namely:

• Bayesian Norm Detection in the bayesian_norm_detector class
• Oren and Meneguzzi's algorithms from COIN
  • A basic norm detector in the basic_norm_detector class
  • A threshold norm detector in the treshold_norm_detector class

Benchmarks were originally implemented as stand-alone scripts, including aamas_experiments.py and journal_experiments.py