I always encouraged students to write their own RDR engine.
The reason is that the core algorithms are extremely simple and
so they will have more control over their project if they have
implemented it as well as the other infrastructure required
for their project. However, there are are number of versions
available - and probably more than this.
page maintained by Paul Compton. last modified 31/1/2008
Knowledge Systems (PKS) has an RDR product LabWizard.
PKS's main market is Clinical Biochemistry, but
Labwizard is a general RDR tool. It is very high
performance with the inference server able to handle multiple knowledge
bases concurrently and high volume cases. It is based on
Multiple Classification RDR (MCRDR). Labwizard is a commercial
product, but research licences are possible. (Note: I
have a small shareholding in PKS)
- I am unaware if other commercial RDR systems are available for research purposes
Suryanto a former PhD student has a Java
implementation available for research purposes.
Documentation is here.
It is also MCRDR based.
Sammut has incorporated single classification RDR into iProlog.
This also contains a version of Brian
Gaines Induct RDR machine learning algorithm.
Induct is a batch learner
the machine learning workbench also includes a version
of Induct RDR