search

friendly / VCDR

Visualizing categorical data with R
0 stars 0 forks source link

\S 5.4.1 primer on loglin models #27

Closed friendly closed 9 years ago

friendly commented 9 years ago

This subsection, borrowed from (now) Chapter 9, isn't the right thing to use here.

Most of the chapter is about mosaic displays, and the need for such a subsection arises because the ideas behind loglin models are used in \S 5.4.2. Is there something better we could use here? In any case, this needs severe editing/revision.

davidjohannesmeyer commented 9 years ago

My idea was to explain the main idea of a log-linear model using the simplest case - with two variables, and the HairEye example used at the beginning of the chapter. This also introduces the usage of loglm(), and the interpretation of the formulas. Then, the example is extended to three variables in Example 5.7, as required by the mosaic plot in Fig. 5.12.

Do you think that more theory is required to get an intuitive sense on what is going on? We could add the log-linear formulas in Table 5.2 to establish the link.

Why is the notation not fitting?

Best David

On 2015-03-20 16:12, Michael Friendly wrote:

This subsection, borrowed from (now) Chapter 9, isn't the right thing to use here.

  • It is in a section on 3-way tables, yet all it discusses is the two-way case, and it just starts with the basics
  • the notation would need to be revised to fit in here.

Most of the chapter is about mosaic displays, and the need for such a subsection arises because the ideas behind loglin models are used in \S 5.4.2. Is there something better we could use here? In any case, this needs severe editing/revision.

— Reply to this email directly or view it on GitHub https://github.com/friendly/VCDR/issues/27.

FH-Prof. Priv.-Doz. Dr. David Meyer Institut für Wirtschaftsinformatik

Fachhochschule Technikum Wien Höchstädtplatz 5, 1200 Wien T: +43 1 333 40 77-394 F: +43 1 333 40 77-99 394 E: david.meyer@technikum-wien.at I: www.technikum-wien.at

friendly commented 9 years ago

I was just flagging this as something that needed more work so that it fits in and gives the reader what you thought was missing. Maybe less theory from first principles, more of some kind of overview. That is sort of what Table 5.2 tries to do. I'll have another look and see if I can come up with something. There was a TODO that speaks to this issue: % \TODO{Add a textbox or text describing the general scheme for translating % among \loglin shorthand, \R model formulas and independence interpretations.}