BoostingPL - Scalable and Parallel Boosting with MapReduce


What is BoostingPL?

  BoostingPL is a scalable and parallel machine learning tools for Boosting
  (What is Boosting? http://en.wikipedia.org/wiki/Boosting).

  This Project is based on this paper:
 
      Indranil Palit and Chandan K. Reddy, "Scalable and Parallel Boosting
      with MapReduce", IEEE Transactions on Knowledge and Data Engineering
      (TKDE), 2012.
   
  If you want to know the theory and demonstration of algorithms in BoostingPL
  , this paper provides references for further reading.

  BoostingPL use the classifiers in WEKA(http://www.cs.waikato.ac.nz/ml/weka/)
  as weak classifiers, so it depends on weka.

  The MapReduce platform we used is Apache Hadoop(http://hadoop.apache.org/)
  , which is an opensource project and is more popular than other 
  implementations of MapReduce. You can deploy it on Amazon EC2 or your own 
  cluster.

  The cgl-mapreduce version for BoostingPL is also vailable from the TKDE paper 
  authors, you can download "CGL-MapReduce-BoostingPL.zip" in "extra/" directory.
  This cgl-mapreduce version depends on these libraries: Weka, twister-0.8, 
  NaradaBrokering-4.2.2. You can download these from source websites.

  BoostingMR( https://github.com/Ranler/boostingMR ): another project which 
  provide an simple web-based interface for hadoop MR jobs.


License:

  BoostingPL is open source software issued under the GNU General Public 
  License (GPLv3). See the LICENSE included in this directory for more 
  information.



Boosting Classifiers:

  At present we have implemented these classifiers:

    * Boosting Classifiers
      - AdaBoost
      - AdaBoostPL
      - LogitBoost
      - LogitBoostPL

    * Weak Classifiers:
      - DecisionStump


A Simple Testing for AdaBoostPL:

https://github.com/Ranler/boostingPL/wiki/A-Simple-Testing-for-AdaBoostPL


Getting Started:

   See http:///



Documentation:

   See http:///



Experimental Results:

   See http:///