Software Downloads Page

Text Topic Classifier

Given any specific topic, this tool allows the automatic classification of text files into relevant and irrelant sets. The package includes source code in Python and is released under the GNU General Public License (GPL). Example text data files (from NCBI's PubMed) and scripts to train and run classifiers are provided in the package.

For more information please refer to the paper Classifying domain-specific text documents containing ambiguous keywords

Download text-topic-classifier-1.0.zip

GCSpeciesSorter:

GCSpeciesSorter is a binary classification package for distinguishing between two or more species based on the GC contents of their DNA or RNA sequences. It includes source code in Python and is released under the GNU General Public License (GPL). Beyond unpacking, there is no special installation step necessary. Python, LIBSVM9 and/or C4.5, and optionally, BLAST are needed to run the scripts. A README file in the package provides more details about running the scripts. The package includes all the input files mentioned in this paper to use as a tutorial, including test sequence files and BLAST database files.

You can read the paper Distinguishing Species Using GC Contents in Mixed DNA or RNA Sequences for more information.

Download GCSpeciesSorter-1.0.tgz or GCSpeciesSorter-1.0.tgz

TimeSleuth:

TimeSleuth's webpage has been moved to SourceForge.net: timesleuth-rule.sourceforge.net

TimeSleuth 3.33: A tool that uses C4.5 and generates temporal rules from sequential data. TimeSleuth can judge the causality/acausality of a set of rules. Includes source code in Java. Read this readme file for more info. The package also contains C4.5's executables: c4.5.exe (tree generator) and c4.5rules.exe (rule generator), compiled for 32 bit Microsoft Windows operating systems. TimeSleuth works at the rule level, so both programs are needed.
Prolog Output: Download this patch file to enable C4.5 to output rules in Prolo. Try c4.5rules.exe with the new "-p 0" or "-p 1" options. You can find papers about This patch here. Note: The C4.5 executables included in TimeSleuth already have this patch.
The complete source code for C4.5 can be found from Ross Quinlan's homepage.
You can find papers about TimeSleuth here.

Download TimeSleuth 3.33

DIPC: Transparent Distributed Shared Memory

DIPC's webpage has been moved to SourceForge.net: dipc-2.sourceforge.net

DIPC (Distributed Inter-Process Communication) allows Linux users to use System V's semaphores, messages, and shared memory over a network. Among other things, it provides Transparent Distributed Shared Memory over a cluster of Linux Machines.

You can find papers about DIPC here.

Download DIPC 2.0 for 2.2.x kernels

Latest beta version:

Download DIPC 2.1 beta 1 for 2.6.x Linux kernels

For 64-bit AMD or Intel CPUs, replace the kernel patch in DIPC 2.1 Beta 1 with this patch file. This kernel patch compiles under 64-bit AMD CPUs, but needs a small change to compile under 64-bit Intel CPUs.

Read this installation guide for more information, including how to make the kernel compile under Intel's 64-bit CPUs.

Search and Problem Solving: The IMA method

The SCSPS java programme randomly generates Constraint Satisfaction Problems (CSP) and solves them using the Iterative Multi-Agent (IMA) method. In spite of all the acronyms, it is a very simple method indeed. You can find papers about the IMA method here.

Download SCSPS.java

Artificial Life

Take a look at URAL. A simple Aritificial Life Simulation using Situation Calculus for learning and planning. It is written in Java.

Download URAL

Dr. Kamran Karimi, P.Eng.