Tutorial: A Module for Linear Regression

Overview

SAS/IML software makes it possible for you to solve mathematical problems or implement new statistical techniques and algorithms. The language is patterned after linear algebra notation. For example, the least-squares formula familiar to statisticians,

$b = (x^'x)^{-1} x^'y$

can be easily translated into the following Interactive Matrix Language statement:

  
    b=inv(x`*x)*x`*y;

This is an example of an assignment statement that uses a built-in function (INV) and operators (transpose and matrix multiplication).

If a statistical method has not been implemented directly in a SAS procedure, you might be able to program it by using IML. Because the operations in IML deal with arrays of numbers rather than with one number at a time, and the most commonly used mathematical and matrix operations are built directly into the language, programs that take hundreds of lines of code in other languages often take only a few lines in IML.

Solving a System of Equations

Top of Page