A description of an AI system, with a demonstration program and Delphi sourcecode.

Commercial Securities Modeler that uses artificial neural networks and genetic algorithms for customizable prediction

Takes a database of cases described by a combination of real and discrete valued attributes, and automatically finds the natural classes in that data. It can be seen as a Naive Bayes classifier where the class node is hidden. [Free]

Supports several inference algorithms and learning algorithms. Allows simulation of static and dynamic networks, including HMMs, IOHMMs, and Kalman filters.

A library of C++ classes for Bayesian filtering. From the Australian Centre for Field Robotics. [C++, MIT license]

A freeware windows-based program that classifies text based on trained material. Designed for automated essay scoring, BETSY can be applied to any text classification task.

Home page of R. Quinlan. FTP links to FOIL (inductive logic programming) and C4.5 (learning decision trees).

A site by Elad Yom-Tov, co-author of the toolbox, that contains additions and updates to the toolbox, as well as a discussion board

Shotaro Akaho's implementation of EM algorithm for modeling Mixtures of Gaussians (Java, free). An extended version is available from the author.

Computational model of human language acquisition written in Java; currently acquires a protolanguage of nouns and verbs language based on visual perception

Generates Gaussian mixture models for large datasets using efficient EM clustering algorithms. [Free]

Hidden Markov Models software library from the Center of Applied Informatics, Cologne. Includes algorithms such as Viterbi, Baum-Welch, and Forward-Backward. [C, GPL license]

On this page an imlementation of Hidden Markov Models and an application to part-of-speech tagging. Also available a multivariate hypothesis testing software for Gaussian Data and TRUEVIZ: A groundtruth/metadata Editing and Visualizing Toolkit for OCR.

Sean Eddy's lab, present profile hidden Markov models for biological sequence analysis, a tool used to build HMMs from multiple alignments, and calculate e-scores.

An algorithm that incrementally constructs decision trees from labeled examples. [Free for individual research purposes]

A general purpose library for Bayesian filtering. [C++, LGPL license]

A software package developed at MIT Lincoln Laboratory which integrates more than 20 neural network, statistical, and machine learning classification, clustering, and feature selection algorithms into a modular software package. [Public domain license]

Links to FTP repositories including ACCEL, CLASSWEB, FOCL, FOIL, GOLEM, INDEX, MILES, MOBAL, OC1, Occamn, PEBLS, RWM.

QM: Guiding inductive learning with a Qualitative Model. LPE: Lazy Partial Evaluation. CN2: Rule induction from examples. [Free]

Software toolkit for building and using motif-based hidden Markov models of DNA and proteins - from the Univ. of California-San Diego.

Software for learning Mixture Distributions. Commercial license.

MEME System is a program for discovering motifs in groups of related DNA or protein sequences. MAST is a tool for searching biological sequence databases for sequences that contain one or more of a group of known motifs.

Suite of Perl tools for counting and analyzing word n-grams in text; provides standard tests of association for identifying word n-grams in large corpora and allows users to implement other tests with minimal Perl knowledge.

Machine learning library for Observable Operator Models (OOMs) suitable for time-series and sequence data classification and prediction. OOMs are similar but more powerful than HMMs. [C++, BSD license]

A C++ library for many pattern recognition tasks; main focus is on image analysis, but a general architecture and XML-based data interchange format allows it to be used for many other tasks as well.

A large collection of multiple sequence alignments and trained hidden Markov models covering many common protein domains.

An architecture for planning and learning. [Free]

A collection of tools for creating and using HMMs for biological sequences. Free license for academic and nonprofit usages.

A learning architecture specifically taylored for learning in very high-dimensional feature spaces. The current release uses sparse variations of Winnow, Perceptron, and Naive Bayes. Free for personal academic and research purposes.

Directory of software tools for modeling graphs and Bayesian networks. Some have learning capabilities.

A program for inducing Bayesian decision trees. Applications to speech. [Free]

The program discovers interesting and repetitive subgraphs in a labeled graph representation using the minimum description length principle. Applications to molecular biology. [Free]

A library of C code useful for writing statistical text analysis, language modeling and information retrieval programs. The current distribution includes the library, as well as front-ends for document classification (rainbow), document retrieval (arrow) and document clustering (crossbow). [Free]

A general approach to the problem of inducing natural language parsers. It uses an annotated corpus, and produces a parser by using ILP for inducing the rules that control the actions of a shift-reduce parser. [Free]

A propositional theory refinement system that will modify a incomplete or incorrect rule base so as to make it consistent with a set of input training examples. [Free]

A program implementing several memory-based learning techniques. These learners store representation of the training set explicitly, and classifies new cases by extrapolation from the most similar stored cases. Free for educational or non-commercial research purposes.

A learning computer player for the light cycles game in Tron

Suite that implements decision trees and tables, rule learners, Naive Bayes, support vector machines, voted perceptrons, multi-layer perceptron. Meta schemes include bagging, stacking, and boosting. [Free under GPL]

Tools for learning dependency networks or Bayesian networks from data. [Free]

Platform- and data-source-independent library for embedded data mining based on the CWM/OMG and other data mining standards. XELOPES-Java algorithms: SVMs, market basket analysis, sequence analysis, decision trees, cluster analysis, multidimensional grouping. XELOPES-C++ algorithms: SVMs, decision trees. [GPL]