Page 319 - Applied Statistics with R

14.2. PREDICTOR TRANSFORMATION 319

95%
-80

log-Likelihood -120 -160

-200

-2 -1 0 1 2

λ

Using the Box-Cox method, we see that = 0 is both in the interval, and
extremely close to the maximum, which suggests a transformation of the form

log( ).

So the Box-Cox method justifies our previous choice of a log transform!

14.2 Predictor Transformation

In addition to transformation of the response variable, we can also consider
transformations of predictor variables. Sometimes these transformations can
help with violation of model assumptions, and other times they can be used to
simply fit a more flexible model.

str(autompg)

## 'data.frame': 383 obs. of 9 variables:
## $ mpg : num 18 15 18 16 17 15 14 14 14 15 ...
## $ cyl : Factor w/ 3 levels "4","6","8": 3 3 3 3 3 3 3 3 3 3 ...
## $ disp : num 307 350 318 304 302 429 454 440 455 390 ...
## $ hp : num 130 165 150 150 140 198 220 215 225 190 ...
## $ wt : num 3504 3693 3436 3433 3449 ...