Dear all,
My question is not at all directly related to Splus,
but perhaps to the statistics. This is in regards to
the option "offset" one can use in fitting log-linear
model for the data which is dependent on the offset by
a direct relationship; perhaps the parameter is fixed
to "1" with it. Which has an interpretation, if I am
not wrong, adjusting the trend observed due the offset
and relation-ship is obvious between the two but one
does not want to model it.
Question: what can one say about any such variable
which is directly related to the dependent variable in
logistic regression.
For e.g. I am using the logistic regresion on the
binary variable (Y) which is threshold indicator for a
proportion (p=t1energy/TotalEnergy) whose denominator
is the total-energy. Now can I use the total-energy as
a independent variable while modelling, it perhaps it
is directly correlated with the variable proportion
but I cannot see an argument which is in
pro or against using the TotalEnergy as one
predictor variable.
Any advice/suggestion will highly welcomed.
Thanks,
Jagadish
__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com
|