Hi,
I have a simple problem where I have two or more predictor variables that
range from 0 to 1 and binary response variable (0 or 1). In the two
variable case, the model to fit with maximum likelihood would simply be:
P(Y=1) = (B1*X1 + B2*X2)/(B1+B2)
or if least squares is to be minimized the model would just be
Y = (B1*X1 + B2*X2)/(B1+B2)
I know that I can write these in nls and other packages and fit using least
squares or maximum likelihood. However, since this is just a weighted
average (a regression with the constraint that all slope coefficients or
weights sum to 1); it seems there should be a simpler method I am not
finding.
Anyone have a quick point to a package/function that will optimze weights in
a weighted average or similarly allow a constraint of all regression
coefficients sum to 1?
Thanks,
Seth
--
View this message in context:
http://r.789695.n4.nabble.com/optimize-weights-for-a-weighted-average-tp3613194p3613194.html
Sent from the R help mailing list archive at Nabble.com.
