Economist and statistician Kattamuri Sarma and his book Predictive Modeling with SAS Enterprise Miner are the source for this week's SAS tip. In a review of the book, SAS user Andrea Wainwright-Zimmerman said "The content is wonderful, clear, and thorough."

If you're using SAS Enterprise Miner, I'd invite you to visit Kattamuri Sarma's author page for bonus content--including additional user reviews and a free chapter. You can also sign up to be one of the first to hear about the release of the new edition of the book.

The following excerpt is from SAS Press author Kattamuri S. Sarma and his book "Predictive Modeling with SAS Enterprise Miner: Practical Solutions for Business Applications" Copyright © 2007, SAS Institute Inc., Cary, North Carolina, USA. ALL RIGHTS RESERVED. (please note that results may vary depending on your version of SAS software)

4.3.9 Selection of the Right-sized Tree

Display 4.7 gives another illustration of the selection of the right-sized tree from a sequence of optimal trees. In the graph shown in Display 4.7, the horizontal axis shows the number of leaves in each of the trees in the optimal sequence of trees. The vertical axis shows the average profit of each tree. From the display it is clear that the optimal tree has seven leaves. That is, a tree with seven leaf nodes has the same average validation profit as any tree with more nodes. Hence the tree with seven leaves is selected by the algorithm (see the vertical line corresponding to seven leaves on the horizontal axis).

Display 4.7 shows the initially constructed maximal tree of size 14, before pruning it to a tree of size 7 using the technique described earlier.