ODD ITEMS AND NOTES

AIDS babies update -- a February 2, 1999 New York Times report on a recent follow-up study of HIV positive mothers and newborns. This discusses results that extend the conclusions we make from our old (Feb 1998) data -- referred to here as the "Thailand study" -- and discusses issues of ethics in clinical trials (as discussed in the Problems of Proportions handout)
Some students have shown interest in learning a little about time series, so I've added here some material on time series autoregressive models -- the central class of time series models. We've run into autoregressive ideas a little bit with the extended Geyser study in Homeworks, and time series problems (e.g., exchange rate forecasting), and this material gives a fair overview of models and modelling technology.
This is not part of the course, just a feeder for those of you who requested some info and leads on this (important) topic, and others who may be interested. I'll be happy to chat with any of you that want to go deeper.
- A short handout introducing autoregressive (AR) models.
- A (much) more extensive (but preliminary draft) manuscript on time series and autoregressive model theory, inference and data analysis.
- For applied work with AR models you can explore the basic ar() function in S-Plus, or
  - download this file of my own (much more extensive) S-Plus AR functions and
  - see how to use them in this AR example S-Plus file which fits and explores the EEG series (from the course Data page) with AR models.
- For much more, check here for some related software (Matlab, Fortran, S-Plus) for more elaborate methods and applications in EEG analysis and elswhere.