CMU Artificial Intelligence Repository
Home INFO Search FAQs Repository Root

Machine Learning Algorithms Implemented in Prolog

areas/learning/systems/learn_pl/

   aq1/       AQ-Prolog: Reimplementation of Michalski's AQ for 
              attribute vectors
   arch1/     ARCH1: Winston's incremental learning procedure.
   arch2/     ARCH2: Winston's incremental learning procedure.
   attdsc/    ATTDSC: Bratko's simple algorithm for 
              attributional descriptions.
   cobweb/    COBWEB: Gennari, Langley, and Fisher's 
              incremental concept formation algorithm.
   discr/     DISCR: Generation of discriminants from 
              derivation trees.
   ebg/       EBG: Explanation Based Generalization
   idt/       IDT: Torgos ID3-like system based on the 
              gain-ratio measure
   invers/    INVERS: Two versions of the two main operators 
              for inverse resolution.
   logic/     Logic procedures that are useful for learning
   multagnt/  MULTAGNT: Brazdil's Simulation of a tutoring 
              setting between two agents
   vs/        Version Spaces

Origin:   

   ftp.gmd.de:/gmd/mlt/ML-Program-Library/ [129.26.8.84]
In 1988 the Special Interest Group on Machine Learning of the German Society for Computer Science (GI e.V.) decided to establish a library of PROLOG implementations of Machine Learning algorithms. The library includes - amongst others - PROLOG implementations of Winston's arch, Becker's AQ-PROLOG, Fisher's COBWEB, Brazdil's generation of discriminations from derivation trees, Quinlan's ID3, FOIL, IDT, substitution matching, explanation based generalization, inverse resolution, and Mitchell's version spaces algorithm. Most of the algorithms are copyleft under the GNU General Public License. They are also available by surface mail from Thomas Hoppe (see address below). Files will be distributed via MS-DOS formated 3.5 inch floppy (double, high and extra-high density), which should be included with your request. You can also get them by sending an email message to Thomas Hoppe. Send additional PROLOG implementations of Machine Learning Algorithms, complaints about them and detected bugs or problems to Thomas Hoppe. Send suggestions and complaints about the ftp library to Werner Emde. The directory ftp.gmd.de:/MachineLearning/ contains additional machine learning publications, data, and software, primarily related to the European ESPRIT projects Machine Learning Toolbox (MLT) and Inductive Logic Programming (ILP), the European Network of Excellence in Machine Learning (MLnet) and the Inductive Logic Programming Pan-European Scientific Network (ILPnet). It includes the source code of Stephen Muggleton's and Cao Feng's GOLEM learning system (in /MachineLearning/ILP/public/software/golem) and a BibTeX file with around 325 entries of articles related to ILP (in /MachineLearning/ILP/public/bib). For more information, send mail to Marcus Luebbe .
Version: 9-FEB-94 Requires: Prolog Ports: All of the algorithms are written in Edinburgh Prolog syntax. Copying: GNU GPL CD-ROM: Prime Time Freeware for AI, Issue 1-1 Contact: Thomas Hoppe (Machine Learning Library) Projektgruppe KIT Technische Universitaet Berlin Franklinstr. 28/29, 10629 Berlin, Germany. Werner Emde (ftp library) Gesellschaft fuer Mathematik und Datenverarbeitung, Bonn Keywords: AQ-Prolog, COBWEB, Derivation Trees, Discriminants, EBG, FOIL, GNU GPL, ID3, IDT, Inverse Resolution, Machine Learning, Mitchell, Prolog!Code, Substitution Matching, Version Spaces, Winston's Arch
Last Web update on Mon Feb 13 10:24:30 1995
AI.Repository@cs.cmu.edu