国际流行开源机器学习和模式识别工具

Machine Learning
  M2K - M2K represents the music-specific set of D2K modules designed to create a Virtual Research Lab (VRL) for MIR/MDL development, prototyping and evaluation. M2K provides the framework for the MIREX (Music Information Retrieval Evaluation eXchange) contest, an annual MIR evaluation. D2K, together with a subsidiary set of modules called T2K (Text-to-Knowledge), provide the basic foundation upon which M2K is being developed. D2K/T2K are the result of a ongoing research and development project of the Automated Learning Group (ALG) at NCSA. M2K License: BSD-Like
  Weka - Weka is a collection of machine learning algorithms for data mining tasks written in the Java programming language. Weka contains tools for data pre-processing, classification, regression, clustering, association rules, and visualization. It is also well-suited for developing new machine learning schemes. The book: Data Mining compliments the Weka Software. License: GNU General Public License (GPL) .
  Marsyas - Marsyas is a software framework for rapid prototyping and experimentation with computer audition applications with specific emphasis on Music Information Retrieval. Marsyas provides a general, extensible and flexible architecture that allows easy experimentation with algorithms and provides fast performance that is useful in developing real time audio analysis tools. A variety of existing building blocks that form the basis of most published algorithms in Computer Audition are already available as part of the package. Marsyas is written in C++ and Java and is actively being developed by George Tzanetakis. License: GNU General Public License (GPL)

  Torch - Torch is a machine learning library written in C++ that works on most Unix/Linux platforms. It can be used to train MLPs, RBFs, HMMs, Gaussian Mixtures, Kmeans, Mixtures of experts, Parzen Windows, KNN, and can be easily extended so that you can add your own machine learning algorithms. Torch is currently developed at IDIAP and is described in the paper Torch : a modular machine learning software library Torch 3 has been successfully tested on Linux, SunOS, FreeBSD, OSF1, Mac OS X and even MS Windows. License: Torch 3 is free, distributed under a BSD license.
  NODElib - Neural Optimization Development Engine library is a programming library for rapidly developing powerful neural network simulations. The code is extremely modular, compact, and robust. It is written in an object oriented manner. All of the library code, example and test program source,w documentation, and supporting text is only on the order of about 20,000 lines, which means that NODElib is extremely compact. NODELib is written in C. License: GNU General Public License (GPL) .
  SVM - this package defines support vector machines (SVMs) for both classification and regression problems. The SVMs can use a wide variety of kernel functions. Optimization of the SVMs is performed by a variation of John Platt's sequential minimal optimization (SMO) algorithm. This version of SMO is generalized for regression, uses kernel caching, and incorporates several heuristics; for these reasons, we refer to the optimization algorithm as SMORCH. SMORCH has been shown to be over an order magnitude faster than SMO, QP, and decomposition. License: GNU General Public License (GPL) .
  LAPACK/BLAS (Linux version available from Intel) for matrix math - The BLAS (Basic Linear Algebra Subprograms) are high quality "building block" routines for performing basic vector and matrix operations. Level 1 BLAS do vector-vector operations, Level 2 BLAS do matrix-vector operations, and Level 3 BLAS do matrix-matrix operations. Because the BLAS are efficient, portable, and widely available, they're commonly used in the development of high quality linear algebra software, LINPACK and LAPACK for example. License: Commercial License

  EMD - an implementation of the Earth Movers Distance. The EMD computes the distance between two distributions, which are represented by signatures. The signatures are sets of weighted features that capture the distributions. The features can be of any type and in any number of dimensions, and are defined by the user. License: unknown.
 BNT - Bayes Net Toolbox for Matlab - supports many types of conditional probability distributions, decision and utility nodes, as well as chance nodes, static and dynamic BNs, many different inference algorithms, several methods for parameter learning, regularization and structure learning. License: GNU Library GPL

 Auditory Toolkbox - a collection of tools that implement several popular auditory models for MATLAB. This toolbox will also be useful to speech and auditory engineers who want to see how the human auditory system represents sounds. License: unknown
  Netlab toolbox - consists of a toolbox of Matlab functions and scripts based on the approach and techniques de