Roger Jang

Introduction

This toolbox (MLT, or Machine Learning Toolbox) provides a number of essential functions for machine learning, especially for data clustering and pattern recognition. The full version of MLT is published by the Terasoft. You can download the complimentary demo version which has same functionality for a limited time.

Documentation

• Within MATLAB:
  • Type "mltDoc" to get a list of functions by categories.
  • Type "mltDoc command" for getting online help. For instance, try "mltDoc nbcTrain" to get the help of training a naive Bayes classifier.
  • Type "doc" and click "Machine Learning Toolbox" to get extensive organized help on the toolbox. (You need to add the toolbox to the search path first.)
• MLT function list by categories
• On-line tutorial on "Data Clustering and Pattern Recognition" using the toolbox, with many examples.

Interesting Demos

After downloading the toolbox, you may want to try some of the interesting demos:

• gmmTrainDemo2dCovType01: Animation of GMM (Gaussian mixture models) training with isotropic covariance matrices
• gmmTrainDemo2dCovType02: Animation of GMM (Gaussian mixture models) training with diagonal covariance matrices
• gmmTrainDemo2dCovType03: Animation of GMM (Gaussian mixture models) training with full covariance matrices
• gmmGrowDemo: Animation of GMM (Gaussian mixture models) growing with center splitting
• hierClusteringPlot: Step-by-step hierarchical clustering

gradientDescentDemo: Interactive demo of gradient descent over the peaks function taylorExpansionDemo: Interactive demo of polynomial fit and taylor expansios lcs and editDistance: Visualize the result of dynamic programming for LCS (longest common subsequence) and ED (edit distance)

Citation

If you are using the toolbox for your research, please kindly give reference as follows:

Jyh-Shing Roger Jang, "Machine Learning Toolbox", http://mirlab.org/jang/matlab/toolbox/machineLearning.