jbjb
Revisions to AIRE 430:
Fig 1 was unclear.  To some extent, the unclear nature of Fig 1  that was the point of Fig 1, Decision trees can be very large and hard to read while TAR2 output can be quite succinct.. However, that point can be still be made while giving the graphic more space. Hence, Fig1 is now two figures, one of which has been given a whole page. Further , more support text has been added to explain the new Fig1 and Fig2 better.
End of section1: more added on contributions of this paper and relation to
prior work
Support based pruning: more text added to distinguish it from weighted class
learning.
Definitions at start of section3: clarified and extended.
Start of section 4: explanations of "43%" and "167%" expanded: in particular,
why it is possible to get > 100%
Numerous small typos and strange wordings fixed
Conclusion: some of the excessive claims in that section removed
I would have liked to have put more effort into referee 2's comments. However, 
the referee sheet said that that reviewer had made annotations on the
manuscript. However, those annotations where not sent to me so I
ve done the
best i could without them.
Revisions to AIRE 430:
Trial User
Normal
Trial User
Microsoft Word 10.0
Revisions to AIRE 430:
Title
Microsoft Word Document
NB6W
Word.Document.8