Prev: Aggragate loss distribution
Next: Excelxp Tagsets: is there a way to create multiple sheets on
From: Kevin Viel on 9 Jul 2008 11:51
I posted yesterday code yesterday that can be found here:
http://www.listserv.uga.edu/cgi-bin/wa?A2=ind0807b&L=sas-l&F=&S=&P=8173

I hope that I had no errors, because proof-reading my own code/writing is
doomed. I was surprised by the number of warnings:

"WARNING: The final Hessian matrix is full rank but has at least one
negative eigenvalue. Second-order optimality condition violated."


I did not check the log of my ZIP simulation yet, but it is now on my
list. The using the above code, I obtained the following proportions:

True_p_logit streaminit( 123 ) streaminit( 456 )
seed = 321 seed = 654
-------------------------------------------------
0.125 0.125
0.05 0.131579 0.144737
0.1 0.171053 0.190789
0.15 0.210526 0.197368
0.2 0.171053 0.190789
0.25 0.210526 0.171053
0.3 0.203947 0.171053
0.35 0.125 0.125
0.4 0.118421 0.131579
0.45 0.092105 0.131579
0.5 0.085526 0.092105
0.55 0.072368 0.052632
0.6 0.085526 0.065789
0.65 0.072368 0.098684
0.7 0.098684 0.085526
0.75 0.059211 0.098684
0.8 0.072368 0.105263
0.85 0.072368 0.131579
0.9 0.111842 0.171053
0.95 0.151316 0.164474


I note that these *should* be data derived from NB from RAND
( "NEGBINOMIAL" , p , k ) that are zero-inflated with the above
probabilities.

If anyone has 9.2 and COUNTREG available. I would be interested to see the
results. Otherwise, I appreciate any criticism and corrections.

Kind regards,

Kevin
 | 
Pages: 1
Prev: Aggragate loss distribution
Next: Excelxp Tagsets: is there a way to create multiple sheets on