To Do List

[ Intro ]

Up: Documents ]

---------
To Do
---------

Create first draft of AIRG 2009-04-09 talk

Compare original BDI questions with those used in our survey.
NOTE: I thought we were interested in the Epworth questionnaire but my 
notes say BDI and that is the paper we found.

Make new data set containing summary data

Create a model for each BDI class and use meta learning to combine

Look into techniques for skewed target attributes

Learn Mat Lab

Read Shiven Thesis

Visualization (?)

Treat MS Office *.*x files as binary on Kapow

Make data sets without epworth_score removed

--------
Done
--------

2009-03-29 Complete NIH human subjects certification at 
http://phrp.nihtraining.com

2009-02-25 Fixed epworth_score range in BDI Results Update 2009-02-20 
presentation

2009-02-25 Add ZeroR to regression experimenter configuration

2009-02-21 Find original BDI paper -Thanks Sergio

2009-02-19 Update To Do List

Prepare 2009-02-20 BDI Result Update presentation

Run Linear regression to predict BDI

Use 2 BDI Bins, 0-9, 10+

Look at macro data set and Shivin's thesis

Find a good discretization for BDI, try Entropy based filter from Weka
 - Found medical discretization consistent with filter results


 

by: Keith A. Pray
Last Modified: March 30, 2009 2:29 PM
© 2009 - 2007 Keith A. Pray.
All rights reserved.