Page 131 - Statistics II for Dummies

Chapter 7

Getting Ahead of the Learning

Curve with Nonlinear Regression

In This Chapter
▶ Getting a feel for nonlinear regression
▶ Making use of scatterplots
▶ Fitting a polynomial to your data set
▶ Exploring exponential models to fit your data

n Stats I, you concentrate on the simple linear regression model, where you
Ilook for one quantitative variable, x, that you can use to make a good esti-
mate of another quantitative variable, y, using a straight line. The examples
you look at in Stats I fall right in line with this kind of model, such as using
height to estimate weight or using study time to estimate exam score. (For
more information and examples for using simple linear regression models,
see Chapter 4.)

But not all situations fall into the straight line category. Take gas mileage and
speed, for example. At low speeds, gas mileage is lower, and at high speeds,
gas mileage is lower; but at medium speeds, gas mileage is higher. This low-
high-low relationship between speed and gas mileage represents a curved
relationship. Relationships that don’t resemble straight lines are called
nonlinear relationships (clever, huh?). Looked at simply, nonlinear regression
takes the stage when you want to predict some quantitative variable (y) by
using another quantitative variable (x) but the pattern you see in the data
collected resembles a curve, not a straight line.
In this chapter, you see how to make your way around the curved road of
data that leads to nonlinear regression models. The good news is twofold:
You can use many of the same techniques you use for regular regression, and
in the end, Minitab does the analysis for you.

12_466469-ch07.indd 115 7/24/09 9:39:07 AM