The Idea is to provide libraries for Classification and Prediction algorithms already published by the academic community. The span extends to not only the basic algorithms, but also the recent advancements made in KDD algorithms, for example SWIFT, BOAT and RainForest. The libraries must be scalable and not tied to any one data set.
The Project will also allow, experimentation and formulation of further optimisation of existing algorithms, an in-depth study of the algorithms and the knowledge extraction portion of the KDD process.
We plan to implement the libraries in C++. The GNU C++ Compiler will be used. We would like to see the algorithms deployed on *Nix and Windows platforms.
As a more long-term goal, the project will spawn to include applications for the libraries that we will implement. The application of KDD in messaging infrastructure in general, is an interesting a relatively unexplored area.
The participants are all from the Rochester Institute of Technology and vocieferously support the concept of 'Free' software and other idealogical variants.
The Project is covered under the Gnu Public License.