RULEX Rule Extraction Downloads
The software on this page is copyrighted by Robert Andrews and Shlomo Geva and is provided free of charge for academic use only.
Last updated 21st October 1999.
Purpose of the Page
The aim of the page is to make available the most recent versions of the RBP neural network and the RULEX rule extraction programs. You can find C source code and executables for PC's as well as some sample datasets, weights files and rules files.
Files
- READMEList of files in this release and instructions for RBP, MCRBP and RULEX.
RBP Neural Network
- rbp.crbp source code
- rbp.hrbp header file
- rbp.exerbp executable for PC's
MCRBP Neural Network
- mcrbp.cmcrbp (multi class rbp) source code
- mcrbp.hmcrbp header file
- mcrbp.exemcrbp executable for PC's
RULEX Rule Extraction Progra
- rulex.crulex source code
- rulex.hrulex header file
- rulex.exerulex executable for PC's
MONK1 Dataset
- monk.netNetwork initialisation file for the MONK1 problem
- monk.patTraining data for the MONK1 problem
- monk.tstTest data for the MONK1 problem
- monk.desRulex description file for the MONK1 problem/LI>
- monk.wtsSample trained RBP weights file for the MONK1 problem
- monk.rulesRule file extracted from monk1.wts
IRIS Dataset
- iris.netNetwork initialisation file for the IRIS problem
- iris.patTraining data for the IRIS problem
- iris.tstTest data for the IRIS problem
- iris.desRulex description file for the IRIS problem/LI>
- iris.wtsSample trained RBP weights file for the IRIS problem
- iris.rulesRule file extracted from iris.wts
PIMA INDIAN DIABETES Dataset
- pima.netNetwork initialisation file for the PIMA problem
- pima.patTraining data for the PIMA problem
- pima.tstTest data for the PIMA problem
- pima.desRulex description file for the PIMA problem/LI>
- pima.wtsSample trained RBP weights file for the PIMA problem
- pima.rulesRule file extracted from pima.wts
This document and those it immediately links to under the ~robert
directory are maintained by Robert Andrews.
Your
comments or suggestions will be valued and appreciated
(use robert@fit.qut.edu.au if your client does not support
mailto URLs).