2

I came across the data mining technique AdaBoost but i can not find much information regarding how it works or any examples i can go through, can someone please shed some light in this area?

Also i would like to give prediction and exploration of data a try, does anyone have any examples?

godzilla
  • 3,005
  • 7
  • 44
  • 60

1 Answers1

4

This tutorial explains the idea very well. At the end of the tutorial, in section Problems, item 4 points you to a simple classification problem which you can try, and item 5 points you to the code of the implementation of the Viola-Jones algorithm in the OpenCV library. Its object detection framework employs a variant of the AdaBoost to both select the best features and to train classifiers that use them.

greeness
  • 15,956
  • 5
  • 50
  • 80
  • forgive me for my ignorance, but how does this algorithm differ from say neural networks which also has a set or weights, an error calculation and recalculation of weights? – godzilla Oct 25 '12 at 20:28
  • Boosting is an ensemble learning method. Usually it is constructed at a higher level after you have some underlying classifiers working. If the underlying classiers are linear networks, then AdaBoost builds multilayer perceptrons one node at a time. However, underlying classier can be anything. Decision trees, neural networks, hidden Markov models. – greeness Oct 25 '12 at 20:50
  • do you have a simple example i can walk through? – godzilla Oct 26 '12 at 06:48
  • 1
    how about this video? There is a simple example in it (starting at 20:01) http://videolectures.net/mlss05us_schapire_b/ – greeness Oct 26 '12 at 07:09