Displaying 1 result from an estimated 1 matches for "spaller".
Did you mean:
smaller
2001 Dec 14
1
Logistic regression : dicrepancies between glm and nls ?
...owed that :
a) the alpha estimators can be much more different than they are in the
present example ;
b) the "biases" (differences between estimators) do *not* depend of
initial values choosen for the nls estimation ;
c) they depend on the size of the sample (the larger the sample, the
spaller the "biases") ;
d) the standard error estimates given by glm are lrger than those given
by nls.
Can someone explain to me why those two methods of fitting a (quite)
simple model give so different results ?
I *think* that two methods should end um with the same estimators (at
least as...