questions-notes-2016-06-06-ch7.txt

7.1 Notes:

Hierarchy of linear models through nonlinear models:

* polynomials
* step functions
* splines
* local regression
* generalized additive models

First, we review polynomial regression

Then looking at step functions.  Introduce dummy variables for
the various step ranges.  Use "cutpoints" or "knots" for an
ordered factor variable (e.g. age ranges with cutpoints 18, 25, 40, 65, 90)


7.2 Next.

7.R

What are orthogonal polynomials?  I don't get this.  What does it mean to
"change the basis for the polynomial"?

What is ANOVA?  How do I use it to compare fits?
* See anova(fita,fitb,fitc,fitd) and compare to e.g. Boston data medv~lstat,etc