Class 19: Interactive Regression Scatterplots and Intro to Smoothing
- Primitive animation
- Graphics objects
Regression scatterplot object
- Lisp script for today's class
- Class19.lsp
- Design
- The regression scatterplot object illustrates how to build special
purpose windows from the basic scatterplot objects included in Lisp-Stat.
This object does the following:
- Shows regression lines and smooths;
- Adds special menu items;
- Adds special mouse modes.
- Introduction to Smoothing
- Many smoothing methods share a common foundation that allows them to be
described in terms of a penalized goodness of fit criterion. Although we will
focus on smoothing splines, the ideas are more general and apply, for example,
to Bayesian methods as well. All of the common smoothers also have two crucial
components:
- A linear projection method
- A nonlinear smoothing criterion
We will not get too far with these today, but will try to set up the basic
results.
Next time
Next time we'll dwell on smoothing, covering methods for studying the linear
aspects of the smoothing procedures. We will also look at criteria for setting the
smoothing constant and problems that all such procedures encounter.