AutoUniv (AU) is a tool for creating classification models and generating data sets from them.
AU is freely available for use by the Machine Learning and Data Mining communities in developing and testing algorithms which learn to classify.
With AU you can create models having:
- up to 1000 attributes and ten classes
- discrete and continous attributes
- up to 3000 rules for a model
- varying degrees of specialisation within rule conditions
- attribute factorisation
- relevant, redundant and pure noise attributes
- varying noise levels and lift
- rare classes
- concept and population drift from other models
Developed by Ray J. Hickey
Contact: ray.j.hickey@gmail.com
AU was implemented in Win-Prolog, v4.9 from Logic Programming Associates