Hi everyone,
I have SAS 9.2.
This is more of a math side question hope someone has an answer.
I'm working with acceptance sampling plans specifically double sampling by variables.
So I have nonlinear function binomial pdf and which is iterated and calculated according to a array of probabilities [0..1].
I would like to equalize (I mean not exact match, equivalent) the result of my binomial pdf to bivariate normal distribution to find my 3 decision variables which are used to calculate bivariate function. As an example:
do p=0.0001 to 0.9999 by 0.0001;
Paatt = probbnml(p,n,c);
Zp = probit(1-p);
Pa1interup=sqrt(n1)*(Zp-k1);
Pa2interinup=sqrt(n1)*(Zp-k2);
Pa2interindwn=sqrt(n1)*(Zp-k1);
Pa2interoutup=sqrt(n1+n2)*(Zp-k3);
PA1 = probnorm(sqrt(n1)*(Zp-k1));
PAa = probbnrm(sqrt(n1+n2)*(Zp-k3), sqrt(n1)*(Zp-k2), n1/(n1+n2));
PAb = probbnrm(sqrt(n1+n2)*(Zp-k3), sqrt(n1)*(Zp-k1), n1/(n1+n2));
PA2 = PAa - PAb;
PA = PA1 + PA2;
end;
I would like to make the PA equivalent to Pattr by changing k1,k2,k3 for the whole nonlinear function. So its more of a approximation of k values from the result of Pattr vector. Any information may be helpful.
Thanks
