Rough draft / proposals for final projects in 402:
-------------------------------------------------

Alexis Lubensky

   Economic Catch-up of Countries in Latin America: Analysis of GDP
   and Other Macroeconomic Time Series for 25 Countries 

   Nice start on clustering series by correlation measures, comparing
   Fourier vs ARMA structure, fitting common models to take out common
   features, etc.


Katie O'Leary

   What Qualities Make a Winning Basketball Team?  A logistic
   regression analysis of the 1997--1998 Men's NCAA Division I Teams.

   Nice start.  Lots of effort into collecting and collating the data
   so it will be useful for a logistic regression of the form
    (wins, losses) ~ whatever

   (note: teams aren't really indep binomial, but it may not matter;
   I've asked her to think about assessing this)


Huadong Wu

   Time Series Analysis for Sound Source Recognition: Pattern 
   Recognition in ARMA Parameter Space

   He's summarized basically what he plans to do.  Will he finish?  I
   doubt it.  But OK...


Ty Douglas

   Analysis of Arrival Traffic at the HUB with a Poisson Regression Model

   Good start, actually nearly done with log linear models and has
   found nothing (which is probably hertening).  recommend also try
   inverse link


Kary Myers

   Logistic Regression and Machine Learning from a Pneumonia Data Base

   Good start; this will lead into her summer work with Cooper,
   Wasserman et al.


Jennifer Fridd
Jessica Kitch

   Testing the Payout Distribution of Video Slot Machines with the
   Parametric Bootstrap

   Good start, looking forward to final product.


Alan Charness

   Calculus Recitation Attendance on FCE Day, as a Function of Large
   Lecture FCE's

   Very nice start, using weighted logistic regression analysis.
   Needs to be more explicit about exploring multicollinearity and
   other anomalies.

   I want to be sure he is doing the weighted logistic regressions right.


Christine Hou
Phoebe Papageorgiou

   The Influence of Five Macroeconomic Time Series on Australia's GDP,
   1959--1995

   Looks like a good proposal for analysis.  In addition please check
   resids from multiple regressions for time series structure either

     - directly
     - or by refitting with lme() and ts structure on the resids


MISSING:

Yoo Jin Chang

   (probably something about diabetes)

   Daily variation in blood sugar levels for insulin dependent
   diabetics

   based on data from 
    http://blanche.polytechnique.fr/www.eark/ftp_site/ML_Repository/diabetes/

   still waiting for interim report

Manoo Halati
Chraig Weiner
Jian Bo Yu

   Relations between and time-dependencies of several daily world
   stock indices, 1995--1997.

   A reasonably good start messing around with detrending and
   deseasoning time series data; also tried fitting some ARIMA(p,1,0)
   models, for p a bit too large (11) but not awfully unreasonable.

   Not likely to get much besides mush out though.