Home > Households, Buildings & Industry > Data Reports > Preliminary End-Use Estimates > Nonlinear Regression

 

Nonlinear Regression Statistical Modeling

The annual amount of energy used for major end-uses was estimated for natural gas, electricity, and fuel oil by use of regression equations. For each equation, the dependent variable was the annual consumption of the given energy source for calendar year 1999. For natural gas, five end-uses are modeled (four specific end uses and an “other” category). The basic natural gas equation is:

Total consumption = space-heating component

+ water-heating component

+ cooking component

+ cooling component

+ other component

The basic equations for electricity and fuel oil are similar in form (but have end uses that are appropriate for those energy sources). The nonlinear regression technique was used and the parameters were estimated by using the nonlinear regression procedure (PROC NLIN) of the SAS system.


 

Specific questions may be addressed to:

Alan Swenson
alan.swenson@eia.doe.gov
Phone: (202) 586-1129
FAX: (202) 586-0018

Release date: October 27, 2003

Page last modified on:

http://www.eia.doe.gov/emeu/cbecs/enduse_consumption/nonlin.html