2
REGRESSION ANALYSIS USING PROC IML AND PROC REG
2.1 INTRODUCTION
We discussed basic regression concepts and least squares theory in Chapter 1. This chapter deals with conducting regression analysis calculations in SAS. We will show the computations by using both Proc IML and Proc Reg. Even though the results from both procedures are identical, using Proc IML allows one to understand the mechanics behind the calculations that were discussed in the previous chapter. Freund and Littell (2000) offer an in-depth coverage of how SAS can be used to conduct regression analysis in SAS. This chapter discusses the basic elements of Proc Reg as it relates to conducting regression analysis.
To illustrate the computations in SAS, we will make use of the investment equation data set provided in Greene (2003). The source of the data is attributed to the Economic Report of the President published by the U.S. Government Printing Office in Washington, D.C. The author’s description of the problem appears on page 21 of his text and is summarized here. The objective is to estimate an investment equation by using GNP (gross national product), and a time trend variable T. Note that T is not part of the original data set but is created in the data step statement in SAS. Initially, we ignore the variables Interest Rate and Inflation Rate since our purpose here is to illustrate how the computations can be carried out using SAS. Additional variables can be incorporated into the analysis with a few minor ...
