I owe some summaries to you all.
First, I had a problem with weird prediction values from a logistic
regression model created by glm(). The values should have been between
0 and 1, but I was getting things like -3.47 and 2.9. I had many
responses to that one, you all know who you are. I knew I had to be
missing something very simple and I was. In the call to predict.glm(),
one must specify type, where type = "link" is the default. I don't know
what "link" is, but for probability predictions one must specify type =
"response". I then got what I needed - thank you all so much.
My next question was on the assign() function. It worked like it was
suppose to, but it output "NULL", something I didn't expect. After
Prof. Ripley's constant reminders, I finally remembered to look at the
code for the function, but it led me nowhere. It turns out it is an
S-Plus artifact that when assign() is used alone without a <-, "NULL" is
output by default. No problems, my code was sound. Thanks to Sam
Buttrey, Steve McKinney, and Nick Ellis for pointing that out.
Finally, I would like to wish all of my American counterparts a very
Happy Thanksgiving. Even though we do not know who our next President
will be, we have a great deal to be thankful for. Many countries settle
their leadership disputes with violence and gunfire, causing residents
to fear leaving their own homes. Let our actions during this time of
confusion and uncertainty show what a great and intelligent citizenry we
have by accepting whatever the results are and move on as fellow
Americans.
Win
******************************************************
Winifred C. Lambert, Senior Scientist/Meteorologist
ENSCO Inc.
Aerospace Sciences and Engineering Division
1980 N Atlantic Ave, Suite 230
Cocoa Beach, FL 32931
VOICE: 321.853.8130 FAX: 321.853.8415
lambert.winifred@ensco.com
AMU Quarterly Reports on the Web:
http://technology.ksc.nasa.gov/WWWaccess/AMU/home.html
******************************************************
|