Project details for MLweb

Logo MLweb 1.2

by lauerfab - February 23, 2018, 15:40:27 CET [ Project Homepage BibTeX BibTeX for corresponding Paper Download ]

view (4 today), download ( 0 today ), 0 subscriptions


MLweb is an open source project that aims at bringing machine learning capabilities into web pages and web applications, while maintaining all computations on the client side, i.e., in the browser.

It includes the following.

LALOLib: a javascript library to enable and ease scientific computing within web pages

LALOLib provides functions for

  • linear algebra: basic vector and matrix operations, linear system solvers, matrix factorizations (QR, Cholesky), eigendecomposition, singular value decomposition, conjugate gradient sparse linear system solver, complex numbers/matrices, discrete Fourier transform... ),
  • statistics: sampling from and estimating standard distributions,
  • optimization: steepest descent, BFGS, linear programming (thanks to glpk.js), quadratic programming.

Documentation is available at

See also the benchmark at

ML.js: a javascript library for machine learning

In addition to all the functions of LALOLib, ML.js implements the following algorithms.


  • K-nearest neighbors,
  • Linear/quadratic discriminant analysis,
  • Naive Bayes classifier,
  • Logistic regression,
  • Perceptron,
  • Multi-layer perceptron,
  • Support vector machines,
  • Multi-class support vector machines,
  • Decision trees


  • Least squares,
  • Least absolute devations,
  • K-nearest neighbors,
  • Ridge regression,
  • LASSO,
  • LARS,
  • Orthogonal least squares,
  • Multi-layer perceptron,
  • Kernel ridge regression,
  • Support vector regression,
  • K-LinReg


  • K-means,
  • Spectral clustering

Dimensionality reduction

  • Principal component analysis,
  • Locally linear embedding,
  • Local tangent space alignment

Documentation is available at

LALOLab: a matlab-like development environment

Try it at

Changes to previous version:
  • Add bibtex entry of corresponding Neurocomputing paper
  • Create javascript modules to avoid global scope pollution in web pages
BibTeX Entry: Download
Corresponding Paper BibTeX Entry: Download
Supported Operating Systems: Linux, Windows, Platform Independent, Mac Os X
Data Formats: Ascii, Csv, Libsvm, Json
Tags: Classification, Clustering, Regression, Dimensionality Reduction, Linear Algebra, Development Environment, Scientific Computing, Web
Archive: download here

Other available revisons

Version Changelog Date
  • Add bibtex entry of corresponding Neurocomputing paper
  • Create javascript modules to avoid global scope pollution in web pages
February 23, 2018, 15:40:27
  • Add gaxpy() and documentation on in-place operations
  • Add loo() function to Classifier and Regression models
  • New contributed toolbox for RNN
  • Minor fixes
November 10, 2017, 11:34:49
  • Faster LeastSquares and RidgeRegression with conjugate gradient method
  • LeastSquares now works also with sparse X
  • Faster thin SVD for tall matrices
  • Fix load data file in LALOLab
  • Add examples in LALOLab
July 7, 2017, 14:43:52
  • Add support for complex numbers, vectors and matrices
  • Add basic signal processing (discrete Fourier transform, sound())
  • Add quadratic discriminant analysis
  • Faster Cholesky factorization
June 1, 2017, 11:48:19
  • Optimize use of kernel cache in MSVM.tune()
  • A few other speed-ups (for spectral clustering, eigs, ...)
  • Add colormap() to Lalolab for colormap plots
  • Changes in some examples
  • Minor bug fixes (including plots in IE)
January 17, 2017, 15:47:41
  • Add Logistic Regression
  • Add support for sparse input in fast training of linear SVM
  • Better support for sparse vectors/matrices
  • Fix plot windows in IE
  • Minor bug fixes
June 28, 2016, 16:00:52
  • Improve NaiveBayes classifier
  • Add online training functions for KNN and NaiveBayes
  • Fix save/load workspace in LALOLab
  • Fix nullspace()
  • Small bug fixes
December 17, 2015, 10:29:35
  • Add Regression:AutoReg method
  • Add KernelRidgeRegression tuning function
  • More efficient predictions for KRR, SVM, SVR
  • Add BFGS optimization method
  • Faster QR, SVD and eigendecomposition
  • Better support for sparse vectors and matrices
  • Add linear algebra benchmark at
  • Fix plots in LALOlib/ML.js
  • Fix cross-origin issues in new MLlab()
  • Small bug fixes
October 9, 2015, 11:55:52
  • Smaller source package
  • Fix Makefile
  • Fix MathJax path
September 22, 2015, 09:57:44
  • Changed name of MLlib for ML.js
  • Improved documentation for ML.js
September 15, 2015, 14:06:39


No one has posted any comments yet. Perhaps you'd like to be the first?

Leave a comment

You must be logged in to post comments.