readme.txt

boost算法

 README 
--------

Directory contains the following files.

1. ADABOOST_te.m
2. ADABOOST_tr.m
3. demo.m
4. likelihood2class.m
5. threshold_te.m
6. threshold_tr.m

The aim of the project is to provide a source of the
meta-learning algorithm known as AdaBoost to improve
the performance of the user-defined classifiers.

To make use of adaboost, first two functions must be
run with the appropriate parameters. The explanation
of each source file is available with "help" command.

To see how they work, run demo.m as

>> demo

First three lines in demo.m specifies the training and
testing set size and the number of weak (threshold)
classifiers.

For bug reporting and for comments do not hesitate to
send e-mail to the author.

Cuneyt Mertayak
email: [e-mail]
version: 1.0
date: 03/09/2008