X******E
发帖数: 353 | 1 需要用R to fit a Weighted Least Square model for a binary response variable
Y.
The weight is a function of E(Y). The model is actually the model used for a
binary
dependent variable when the logistic model has not been invented.
我知道当weight is $E(Y)^{-2\alpha}$, varPower() in the package of gls() can
provide an optimal alpha estimator. But what I need is 1/(E(Y)(1-E(Y)). I
know how to do it in SPSS,
but do not know how to do it in R or SAS. Could someone help me out?
Thanks for any replies in advance! |
|
