Displaying 1 result from an estimated 1 matches for "xeta".
Did you mean:
beta
2008 May 19
0
constrOptim converging not to the optimal values
...a function subject to inequality
constraint. It works well when the number of parameters to optimize is low
(e.g. 4) but when they are more (e.g. 10) it does not produce the expected
results.
The function is minus the determinant of a binomial logit model with one
explanatory variable.
Calling ?xeta? the vector of explanatory variables the function to be
optimized is:
negdet<-function(xeta)
{
pi <- exp(a + b* xeta)/(1+exp(a + b* xeta))
-( sum(pi*(1-pi))*sum(pi*(1-pi)*xeta^2) - sum(pi*(1-pi)*xeta)^2 )
}
With ?a? intercept and ?b? parameter of xeta in a standard binomial logit,
i.e. p...