10-21-2014 01:33 PM
I know proc genmod can do logisitc regression to model discrete response (0 or 1). However, what I have now is a continuous binomial distributed response (anywhere between [0,1] but the distribution is binomial with probability p). Is there a way for me to use proc genmod to do logisitic regression on it? thanks.
10-21-2014 02:29 PM
There is no such thing as a continuous binomial distribution. The binomial distribution is discrete. Maybe your data follows a logistic distribution? Or maybe what you want is to fit a (non-linear) logistic function to your data?
Is there an independent variable involved?
10-21-2014 03:35 PM
Yeah. I should say I know the binomial distribution's probability p (which is a continuous ratio) and its total count. For example, I have 100 obs with prob 0.1, 200 obs with prob 0.2, 300 obs with prob 0.3, .... Each obs has some independent variables which can be class or numeric variables. How can I use proc genmod to do a logistic regression on it? Thanks.