Post Reply 
Logistics Rgression in Statistics 2
11-13-2016, 10:00 PM
Post: #1
Logistics Rgression in Statistics 2
The manual indicates that if you store a value in the real variable "L" (other than ZERO) then it will be taken as the "saturation" value in determining the best fit parameters for a logistic function (regression model). This does not seem to work. The system provides its own estimate for "L".
Has anyone a solution or is this a bug.
Find all posts by this user
Quote this message in a reply
11-14-2016, 12:51 PM (This post was last modified: 11-14-2016 01:00 PM by mark4flies.)
Post: #2
RE: Logistics Rgression in Statistics 2
I just confirmed that the procedure will fit the model parameter L. It is not the global variable L, as far as I can tell. I set to a value and refit the data, with the same result as the first time. (Confirming your claim.) How did you try to fix this parameter to a particular value?

I wonder if it checks for L undefined (L=0) only once or every time? L was undefined the first time I fit data.
Find all posts by this user
Quote this message in a reply
11-14-2016, 03:55 PM
Post: #3
RE: Logistics Rgression in Statistics 2
Well, this is the first time I'm hearing about this "feature" and I wrote the logistic regression code so...

Seems like a case of the person writing the documentation including a feature that wasn't included in a note to the developers. Smile

I'll put it on the list and see what can be done.

TW

Although I work for HP, the views and opinions I post here are my own.
Find all posts by this user
Quote this message in a reply
11-14-2016, 11:26 PM
Post: #4
RE: Logistics Rgression in Statistics 2
Tim,
Thanks. It would make the function more useful from a teaching perspective - biology projects.
Regards
Chris Bragg
Find all posts by this user
Quote this message in a reply
Post Reply 




User(s) browsing this thread: 1 Guest(s)