ml - Machine Learning Package

  What is it?
  -----------

  The ml package contains a list of libraries that can be used for supervised
  learning. All the libraries are developed in Java. It was developed by me
  (Harihara Narayanan) as part of my grad class in Machine learning. The main
  purpose of this package is to enable people to use various classifiers without
  worrying about the internal details.  

  Contents
  ----------
  
  This package will be constantly updated to include libraries for new classifers
  (as and when I learn them). At the time of last update - it contains the following
  libraries. I have provided a very concise definition of each classifier - you would be 
  best advised to refer to tutorials on these, in case you are just starting out. 

       ID3
      ------ 
      This is the Iterative Dichotomiser library for creating decision trees from
      the given dataset. It operates on the concept of Information Gain, which
      determines the best split attribute based on change in entropy before and
      after splitting. 

      NeuralNetworks
     ----------------
      Neural Networks are software implementation of neurons and their interconnection. 
      It learns using a technique called backpropagation - that uses gradient descent to
      identify a state when the Mean squared error of the outputs is at a minimum. 

      KNearestNeighbor
     -----------------
      This is a lazy learning approach that operates on the belief that instances that are
      close to each other have similar class labels.
                        

  Documentation
  -------------

  The documentation for the libraries can be found in this repository's wiki pages.

  Licensing
  ---------

  This work is protected by the DBAD license. Refer to http://philsturgeon.co.uk/code/dbad-license. 

  Contacts
  --------

  For any questions, comments or bug reporting, contact me at hariharakrishnan.n@gmail.com.