[marf-cvs] marf/src/marf/FeatureExtraction/RandomFeatureExtraction RandomFeatureExtraction.java, 1.15, 1.16

[Serguei A. M. <mo...@us...>]

Update of /cvsroot/marf/marf/src/marf/FeatureExtraction/RandomFeatureExtraction
In directory sc8-pr-cvs7.sourceforge.net:/tmp/cvs-serv4771/marf/src/marf/FeatureExtraction/RandomFeatureExtraction

Modified Files:
	RandomFeatureExtraction.java 
Log Message:
Massage Classification such that the decendants can still use 
properly the generic dump()/restore() API. Correct the bug
in Classification #1515258 when restore() was calling super.dump()
in the default: case. Fix a similar bug in NeuralNetwork.
Implement and test the work of DUMP_GZIP_BINARY and DUMP_BINARY modes
for the NeuralNetwork and DUMP_BINARY for the TrainingSet. In the
NeuralNetwork refactor the network data structure generation code
into the generate() method so it can be used with both binary and XML
(and other) dump modes. The changes to Classification, NeuralNetwork,
and TrainingSet together resolve the bug #1508630.
Improve TrainingSet's implementation to properly use the general
StorageManager's API of dump()/restore().
Make more use of StringBuffer vs. String concatenation to improve performance
and avoid memory exhaustion.
Make RandomClassificaion to use the general dump()/restore() API of the
StorageManager instead of its own.
Improve error reporting.
Add/massage some comments along the way.


Index: RandomFeatureExtraction.java
===================================================================
RCS file: /cvsroot/marf/marf/src/marf/FeatureExtraction/RandomFeatureExtraction/RandomFeatureExtraction.java,v
retrieving revision 1.15
retrieving revision 1.16
diff -C2 -d -r1.15 -r1.16
*** RandomFeatureExtraction.java	21 Jan 2006 02:35:32 -0000	1.15
--- RandomFeatureExtraction.java	28 Jul 2006 17:13:02 -0000	1.16
***************
*** 12,16 ****
   * <p>Implementation of random feature extraction for testing as a baseline.</p>
   *
!  * <p>$Id$<p>
   *
   * @author Serguei Mokhov
--- 12,16 ----
   * <p>Implementation of random feature extraction for testing as a baseline.</p>
   *
!  * $Id$
   *
   * @author Serguei Mokhov
***************
*** 33,37 ****
  	 * @since 0.3.0.4
  	 */
! 	private static final long serialVersionUID =-5469714962808143269L;
  
  	/**
--- 33,37 ----
  	 * @since 0.3.0.4
  	 */
! 	private static final long serialVersionUID = -5469714962808143269L;
  
  	/**
***************
*** 49,53 ****
  	 * Random Gaussian-based feature extracton.
  	 * @return <code>true</code> if successful
! 	 * @throws FeatureExtractionException
  	 */
  	public final boolean extractFeatures()
--- 49,53 ----
  	 * Random Gaussian-based feature extracton.
  	 * @return <code>true</code> if successful
! 	 * @throws FeatureExtractionException in case of any error
  	 */
  	public final boolean extractFeatures()
***************
*** 82,85 ****
--- 82,86 ----
  		catch(Exception e)
  		{
+ 			e.printStackTrace(System.err);
  			throw new FeatureExtractionException(e);
  		}