Graph-RAT is a development environment for bridging the gap between data mining, recommendation, and machine learning. At its core, it is a database abstraction layer, abstracting away differences between XML files, database files, and multiple database sources. On top of this is a query language that encompasses all SQL queries that do not require temporary storage (with guarantees of constant space iterators). On top of the query system is a language of analysis algorithms that function as analysis primitives. These algorithms are controlled by a scheduler (by default, sequential execution of algorithms.)

The benefit of this complexity is the ability to mingle nearly all different approaches to data mining in a single system. In addition, certain kinds of analysis are only possible in this system. For instance, defining subgroups, performing analysis of the real-time-defined subgroups, then aggregating the results into a global result.

The syntax of the language is a simple XML format. However, Graph-RAT is designed to utilize GUI tools to write programs - the tree structure of the analysis is poorly suited to linear text descriptions. Graph-RAT 0.4.3 provides such a GUI, while Graph-RAT 0.5.1 is not yet implemented.

The current stable branch is 0.4.3, containing most of the analysis algorithms needed for building software. The current development branch is 0.5.1 - a complete rewrite of the system adding a query system, a more complete metadata standard for describing algorithms, and a more robust and user-friendly system for describing parameters and properties. 0.5.1 is in alpha, currently experiencing stability issues with only embedded, not GUI or command-line functioning yet. This project has been a target for hackers damaging some source. However, a code review has been performed and there are no hacker additions to the code itself though older binaries should be scanned before use.

• Using an individual algorithm: see the algorithm Javadoc for the citation.
• Using as a language: McEnnis, Daniel. 2009. "Graph-RAT programming environment" Waikato University White Paper. 04/2009.
• Using Graph-Triples: McEnnis, Daniel. 2009. "Linear-time graph triples census algorithm under assumptions typical of social networks." University of Waikato White Paper. 06/2009.