27.12.2013 Views

089-2013: Using Classification and Regression Trees (CART ... - SAS

089-2013: Using Classification and Regression Trees (CART ... - SAS

089-2013: Using Classification and Regression Trees (CART ... - SAS

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>SAS</strong> Global Forum <strong>2013</strong><br />

Data Mining <strong>and</strong> Text Analytics<br />

<strong>Using</strong> <strong>Classification</strong> <strong>and</strong> <strong>Regression</strong> <strong>Trees</strong> (<strong>CART</strong>) in <strong>SAS</strong>® Enterprise Miner TM , continued<br />

CONTACT INFORMATION<br />

Your comments <strong>and</strong> questions are valued <strong>and</strong> encouraged. Contact the author at:<br />

Name: Leonard Gordon<br />

Enterprise: University of Kentucky<br />

Address: 725 Rose Street<br />

City, State ZIP: Lexington, KY 40536<br />

Work Phone: 859-218-2097<br />

Fax: 859-257-6430<br />

E-mail: leonard.gordon@uky.edu<br />

<strong>SAS</strong> <strong>and</strong> all other <strong>SAS</strong> Institute Inc. product or service names are registered trademarks or trademarks of <strong>SAS</strong><br />

Institute Inc. in the USA <strong>and</strong> other countries. ® indicates USA registration.<br />

Other br<strong>and</strong> <strong>and</strong> product names are trademarks of their respective companies.<br />

APPENDIX<br />

These are the English rules for the classification tree<br />

*------------------------------------------------------------*<br />

Node = 3<br />

*------------------------------------------------------------*<br />

if respillness IS ONE OF: 0 or MISSING<br />

then<br />

Tree Node Identifier = 3<br />

Number of Observations = 407<br />

Predicted: asthma=0 = 0.95<br />

Predicted: asthma=1 = 0.05<br />

*------------------------------------------------------------*<br />

Node = 5<br />

*------------------------------------------------------------*<br />

if respillness IS ONE OF: 1<br />

AND Replacement: cafocoeffecient >= 329716<br />

then<br />

Tree Node Identifier = 5<br />

Number of Observations = 6<br />

Predicted: asthma=0 = 0.00<br />

Predicted: asthma=1 = 1.00<br />

*------------------------------------------------------------*<br />

Node = 13<br />

*------------------------------------------------------------*<br />

if respillness IS ONE OF: 1<br />

AND prematurebirth IS ONE OF: 0<br />

AND Replacement: cafocoeffecient < 329716 or MISSING<br />

AND GENDER IS ONE OF: 1<br />

then<br />

Tree Node Identifier = 13<br />

Number of Observations = 6<br />

Predicted: asthma=0 = 0.83<br />

Predicted: asthma=1 = 0.17<br />

*------------------------------------------------------------*<br />

Node = 15<br />

*------------------------------------------------------------*<br />

if respillness IS ONE OF: 1<br />

AND prematurebirth IS ONE OF: 1 or MISSING<br />

AND anwork IS ONE OF: 1 or MISSING<br />

7

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!