Open Source Java Mathematics Software

UnBBayes is a probabilistic network framework written in Java. It has both a GUI and an API with inference, sampling, learning and evaluation. It supports Bayesian networks, influence diagrams, MSBN, OOBN, HBN, MEBN/PR-OWL, PRM, structure, parameter and incremental learning. Please, visit our wiki (https://sourceforge.net/p/unbbayes/wiki/Home/) for more information. Check out the license section (https://sourceforge.net/p/unbbayes/wiki/License/) for our licensing policy.

LTI-Lib is an object oriented computer vision library written in C++ for Windows/MS-VC++ and Linux/gcc. It provides lots of functionality to solve mathematical problems, many image processing algorithms, some classification tools and much more...

Prenzl!! is a desktop application and a screen saver running artistic cellular automata. The application is based on an Eclipse architecture and the rules can be in native. Our mission is to provide a viewer and an SDK for this kind of art, called CArt.

A java based neural network framework. The Auratus network is built around an XML messaging system, allowing for a complete MVC design. Additionally, Auratus networks are constructed and at the node/edge level, allowing for advanced topologies.

BabyTALK is to add another brick in the wall of natural languages learning. The baby needs to structure a corpus of texts when his tutor points and talks about a particular part of the corpus. The baby is also to describe any selected part of the corpus.

Java/XML toolkit for research using Bayesian networks and other graphical models of probability (exact and approximate inference, structure learning, etc.)

Bayesian Network tools in Java (BNJ) is an open-source suite of software tools for research and development using graphical models of probability. It is published by the Kansas State University Laboratory for Knowledge Discovery in Databases (KDD).

Belief merging aims at combining several pieces of (possibly inconsistent) information coming from different sources. The goal is to produce a single consistent set of information, trying to keep the most of the information of the sources. A belief merging operator is the responsible for making the belief merging. With this tool you can compare the ΔΣ, ΔGMax and the Δps (PS-Merge) operators.

Belief revision is a central topic in knowledge representation and reasoning. It consists in incorporating a new belief, changing as few as possible of the original beliefs while preserving consistency. Revision always considers new evidence as a better belief. Such new evidence is usually represented in the form of a propositional formula which must be preserved after the revision. Here, the Δps (PS-Merge) belief merging operator is extended in order to consider constraints, and this extension is used as a strategy for belief revision. The new evidence is treated as a constraint so the extended merging operator can obtain the revised belief base.

===== Project moved to https://github.com/fpetitjean/Chordalysis ===== Log-linear analysis is the statistical method used to capture multi-way relationships between variables. However, due to its exponential nature, previous approaches did not allow scale-up to more than a dozen variables. We present here Chordalysis, a log-linear analysis method for big data. Chordalysis exploits recent discoveries in graph theory by representing complex models as compositions of triangular structures, also known as chordal graphs. Chordalysis makes it possible to discover the structure of datasets with thousands of variables on a standard desktop computer. Associated papers at ICDM 2013, ICDM 2014 and SDM 2015 can be found at http://www.francois-petitjean.com/Research/ YourKit is supporting Chordalysis open source project with its full-featured Java Profiler. YourKit is the creator of innovative and intelligent tools for profiling Java and .NET applications. http://www.yourkit.com

Fuzzy numbers and basic fuzzy arithmetics (+, -, *, /, 1/x) implementation written in Java. Should be used as a library, but it also has a GUI which demonstrates its capabilities (using SWT widgets).

Fuzzy logic add-in for OpenOffice.org Calc. InrecoLAN FuzzyMath allows to perform ordinary arithmetic operations and use ordinary mathematical and financial functions with fuzzy numbers. Have any ideas how to improve this project - you are welcome!

The K.U.Leuven JCHR System is an integration of Constraint Handling Rules (CHR) and Java designed with three aims in mind: user-friendliness, flexibility and efficiency. http://www.cs.kuleuven.be/~petervw/JCHR/

JECL, pronounced "Jekyll", is an open-source Java library implementing a variety of evolutionary algorithms. The goal is to provide a uniform interface for describing problems, running evolutionary algorithms, and accumulating results. The Java Evolutionary Computation Library is no longer under active development. We suggest the MOEA Framework (http://www.moeaframework.org/), jMetal (http://jmetal.sourceforge.net/), ECJ (http://cs.gmu.edu/~eclab/projects/ecj/) or JGAP (http://jgap.sourceforge.net/) as alternatives.

Java Kohonen Neural Network Library Kohonen neural network library is a set of classes and functions for design, train and use Kohonen network (self organizing map).

Java Metaheuristics (JMH) is a Java library aimed at the design and implementation of exact and approximated algorithms for optimization problems. JMH is specifically tailored for the design of metaheuristic procedures.

Use a language CycL or KIF to represent shared ontological integration with a first person MOO system to provide debugging and agent facilities.

An Java artificial intelligence platform comprising common AI tools (neural networks, genetic algorithms, wavelets, etc...) and examples. The platform focuses on ease of use and ease of integration..

Molle is a cross-platform prover for modal logic, which exploits the modal semantic tableaux method. It features a very usable graphical interface, with interactive representation of generated models. Molle is maintained by Politecnico di Milano.

3-layer neural network for regression and classification with sigmoid activation function and command line interface similar to LibSVM. Quick Start: "java -jar nen.jar"

RISO: distributed, heterogeneous Bayesian belief networks. Belief network: a probability model defined on an acyclic directed graph; distributed: nodes can be on different hosts; and heterogeneous: allowing different types of conditional distributions.

The Simple Interface for Global Optimization Algorithms allows the specification of arbitrary search/optimization problems, solving of these problems, and the specification and implementation of optimization algorithms like evolutionary algorithms.

SNARLI (Simple Neural ARchitecture LIbrary) is a Java package containing a back-prop layer class and classes for the Self-Organizing Map and Incremental Growing Grid. The back-prop class supports sigma-pi connections and back-prop-through-time.

The Word Problem Solver is a program that is able to solve math word problems like they are teached in the first four classes in primary school. Currenty the program is restricted to math word problems that are presented in german language.

mbFuzzIT enables the work with Fuzzy technology (fuzzy inference / fuzzy logic) in Java. In order to facilitate the creation of the inference mechanisms, mbFuzzIT containts a visual editor.