JLinAlg

This is the one and only homepage of the JLinAlg-project,
an open source Java-library for linear algebra that is licensed
under the GNU General Public License (GPL).



		[Features] [Download] [Documentation] [Examples] [Project Members] [Links]

Features

Basic matrix and vector operations like scalar product and matrix multiplication.
Matrices and vectors can be filled with arbitrary field elements like
rational numbers, complex numbers or even prime fields.
Compute the inverse or the determinant of a matrix.
Gauss- and Gauss-Jordan algorithm are available.
Calculate a solution or the solution space of a linear equation system.
Calculate the eigenvalues of real matrices.
All operations can be done without rounding errors
(except for numeric datatypes like DoubleWrapper).
Element-wise mathematical and logical operations on vectors and matrices.
L1 and L2 vector norms.
Canonical vector and matrix factory methods: identity matrix,
all ones, all zeros, uniformly distributed noise,
Gaussian (normally distributed) noise.
Application of user-defined functions to all elements of a matrix
or vector (or pairs of matrices and vectors).

Changes in version 0.35: Bug-fixes

Changes in version 0.34:

Added fast matrix multiplication (see MatrixMultiplication - experimental)
Bug-fix in LinSysSolver

Changes in version 0.33:

New FieldElement: BinaryFloat
LinAlgFactory for arbitrary FieldElements.

Download

JLinAlg-v0.35: The complete source code and all javadocs in one zip-file.
Former versions: Other releases at Sourceforge.

Documentation

JLinalg API-Documentation: Quite thorough documentation of every method and constructor.
Project-CVS: Browse the CVS-repository of JLinAlg.

Examples

Linear equation system: This example shows how to calculate a solution or its solution space.
Determinant/Inverse: How to calculate them for matrices over arbitrary fields.
EigenvaluesDemo.java: This example shows how JLinAlg can do eigenvalue computation
with real matrices (matrices with DoubleWrapper entries).
This demonstration of the class FieldP schows that in vector spaces over finite fields there
can be linear dependent vectors which are all orthogonal to each other.( FieldPDemo.java ).
Exclusive-Or neural net problem using JLinAlg ( Xor.java ).

Project Members