Information Technology Reference
In-Depth Information
sas ProgrammIng and data mInIng
In order to define the different indices used in this text, the data sources require some preprocessing.
Some of the processing code will be given here. More details are provided in workbooks developed
for preprocessing the datasets.(Cerrito, 2008a, 2008b) We do assume that the reader either has some
experience in working with SAS software, or collaborates with a user of SAS software to preprocess
the data. For a general introduction to programming SAS using the point-and-click interface, Enterprise
Guide, we refer the reader to Exploratory Data Analysis: An Introduction to Data Analysis Using SAS.
(P. Cerrito, 2007)
We will also be using the SAS component, SAS Enterprise Miner for all data mining. This is a data
mining add-on to SAS software. It, too, has a point-and-click interface similar to that in SAS Enterprise
Guide. We will give a brief introduction to Enterprise Miner as we will be using several of the compo-
nents in this topic. For more information, we refer the interested reader to a text on Enterprise Miner.
(P. B. Cerrito, 2007)
In particular, we will make use of a feature in SAS Enterprise Miner called Text Miner. The purpose
of this component is to analyze unstructured, non-standardized text. It can also be used to examine
patient information that is identified using ICD9 codes. In particular, SAS Text Miner has the ability to
cluster patients into groups that can be ranked by order of severity. Once ranked, they can be validated
by using patient outcomes to determine if the most severe patients have the greatest proportion of det-
rimental outcomes.
Figure 1 shows the login screen for Enterprise Miner. For a desktop install, the password is the Win-
dows login user id and password.
Figure 1. Login for Enterprise Miner (created with SAS® software. Copyright 2009, SAS Institute Inc.,
Cary, NC, USA. All Rights Reserved. Reproduced with permission of SAS Institute Inc., Cary, NC)
Search WWH ::
Custom Search