development.txt

Java方面的数值算法

This files contains some basic information about the development of the 
Java Machine Learning API.

1. Future Milestones
====================
0.1.0 --> the library should have a set of validated clustering algorithms with 
evaluation criteria and datasets.

0.2.0 --> the library should include several implementations of SVMs

0.3.0 --> the library should include several features selection techniques

0.4.0 --> the library should support missing values

0.5.0 --> the library should include several implementations of Neural Nets

1.0.0 --> the library should have a set of validated machine learning algorithms 
including but not limited to: clustering, classifiers, features selection 
(filters). This should at least cover the algorithms that are available in 
WEKA. By this release the interfaces should be stable and backward compatibility
should be maintained from this point forward.


2. Release protocol
===================
The package consists of 9 files and 1 directory

Files:
1. changelog in txt format
2. changelog in html format
3. Java Machine Learning API source files in jar format 
	(net.sf.javaml package and junit package)
4. Java Machine Learning API as a binary library in jar format 
	(net.sf.javaml package)
5. readme.txt
6. quickstart.txt
7. development.txt (this file)
8. license
9. commmons math library

Directories:
1. The javadoc documentation: this is the documenation of the net.sf.javaml 
package. Other packages are excluded.