SASIML

1
SAS/IML Interactive Matrix Language March
23, 2005Presented by theITC Research Computing
Support Group Kathy Gerber, Ed
Hall, Katherine Holcomb, Tim F. Jost Tolson

• Finding Global Optima with the Genetic
  Algorithm/ Direct Search Toolbox, Wednesday,
  April 6, 330 PM
• Statistical Shape Analysis, Wednesday, April 20,
  330 PM
• Objected Oriented Programming with Fortran 95,
  Wednesday, April 27, 330 PM

2
SAS/IML

• SAS Interactive Matrix Language

3
Outline

• Overview of SAS/IML
• Language nuts and bolts sample programs
• IML Workshop 2.1

4
Features of SAS/IML

• Is a programming language.
• Has has a complete set of control statements.
• Operates on matrices.
• Has a powerful vocabulary of operators.
• Can use operators that apply to entire matrices.
• Can be interactive.
• Is dynamic.
• Processes data.
• Can produce graphics.

5
When to Use SAS/IML

• If SAS is currently your major tool for data
  analysis and quantitative research, then SAS/IML
  software as a programming language extends the
  capability of the existing environment.
• When an environment combining existing
  statistical/ mathematical procedures and
  versatile programming capability is required.
• If a calculation is unavailable in SAS/STAT or
  Base SAS, write the core algorithm in an IML
  module.
• When additional flexibility and integration are
  required for tasks such as data acquisition,
  generation and transformation, obtaining and
  reusing results, and regular expression parsing.

6
Caveats

• Matrix operations are not optimized.
• Base SAS, SAS/IML and SAS Macro languages differ
  syntactically.
• Limitations on data structures.

7
With SAS/IML

• You can access SAS data sets or external files
  with an extensive set of data processing commands
  for data input and output
• You can also edit existing SAS data sets or
  create new ones
• You can leverage familiar statistical procedures
  from within your code

8
Numerical Functions and Algorithms Implemented
with SAS/IML

• Many Base SAS functions are accessible from
  SAS/IML
• Subroutines for outlier detection and robust
  regression
• Kalman Filters
• Time Series Analysis
• Wavelet Analysis
• Genetic Algorithms Experimental
• Sparse Matrices Experimental
• Monte Carlo Simulation see http//support.sas.co
  m/publishing/bbu/companion_site/57323.html

9
A Few Programs That Use SAS/IML

• Sample programs installed with your installation
  
• Located in directory C\Program
  Files\SAS\SAS 9.1\iml\sample
• Samples included in SAS Online Documentation for
  SAS/IML
• http//support.sas.com/91doc/
• A collection of SAS macro programs using SAS/IML
  software to generate, randomize and inspect
  orthogonal arrays for computer experiments and
  integration.
• http//sunsite.univie.ac.at/statlib/designs/
  oa.SAS
• SAS for Monte Carlo Studies A Guide for
  Quantitative Researchers By Xitao Fan, Akos
  Felsovalyi, Stephen A. Sivo, and Sean C. Keenan
• http//support.sas.com/publishing/bbu/compa
  nion_site/57323.html
• Example for this talk
• SAS Program http//www.itc.virginia.edu
  /research/talks/sasiml.sas
• Data Set http//www.itc.virginia.edu
  /research/talks/gpa.sas7bdat

Starting SAS program
10
SAS/IML Workshop 2.1

• SAS/IML Workshop is a graphical user interface
  and extension of SAS/IML software.
• Enables you to explore data interactively using
  standard statistical graphics and tables.
• Provides an integrated development environment
  for writing, debugging, and executing IML
  programs.

11
Using IML Workshop, you can

• Analyze data using interactive statistical
  graphics. IML Workshop provides bar charts, box
  plots, contour plots, histograms, line plots,
  mosaic plots, and two- and three-dimensional
  scatter plots.
• Execute SAS procedure steps from within your
  IMLPlus programs. This enables you to extend
  IML's capabilities with the power of SAS'
  analytic procedures.
• Develop interactive data analysis programs using
  dialog boxes.
• Call your own computational routines written in
  C, Fortran, or Java.

12
System Requirements and Download

• SAS/IML Workshop requires a PC running the
  Microsoft Windows operating system - Windows
  98/ME/NT4/2000/XP .
• First install SAS Release 8.2 or Release 9.1
• Sun Microsystems's Java Runtime Environment (JRE)
  version 1.4.1 or later must be installed. You can
  download the latest JRE from Sun's Java Web site
  http//www.java.com.
• If you are using Windows 98, Microsoft DCOM98
  version 1.3 must be installed. You can download
  DCOM98 1.3 from Microsoft's Web site
  http//www.microsoft.com/com/dcom/dcom98/dcom1_3.a
  sp.
• Pentium-compatible processor with a clock speed
  of 233 MHz or higher.
• Memory - A minimum of 64MB RAM is required 128MB
  RAM or more is recommended.
• Hard disk requirements 62MB required to
  install.
• Request download from http//www.sas.com/apps/demo
  sdownloads/setupcat.jsp?catSAS2FIMLSoftware

13
Differences in IMLPlus

• The mechanism for mixing IML with DATA steps and
  other PROC steps is different in IMLPlus. Refer
  to the help topic Accessing SAS Functionality.
• IMLPlus programs do not require the statements
  proc iml or quit.
• The following low-level graphics routines are not
  supported display, gpie, gpiexy, gpoint, gpoly,
  gport, gportpop, gportstk, gscript, gset, gshow,
  gstart, gstop, gstrlen, gtext, gvtext, gwindow,
  gxaxis, gyaxis, and window. In their place
  IMLPlus provides high-level statistical graphics.

Starting IML Workshop
14
Summary

• PROC/IML can be an ideal tool within an existing
  SAS context or for small to midsize independent
  numerical programming projects
• IML Workshop 2.1 is particularly handy for quick
  point-and-click interactive dynamic visualization
  of numerical data
• NOT silver bullets but useful tools