Prev: book: Clean Code
Next: Import data from Excel
From: "Agrimson, Andrew" on 19 Nov 2009 15:45
Hello all,

=20

I'm estimating parameters using the log likelihood of a gamma -gamma
mixture model using NLP. =20

=20

I'm getting the following warning in the log:

=20

=20

NOTE: Your code contains 1 program statements.

NOTE: Gradient is computed using analytic formulas.

NOTE: Hessian is computed using finite difference approximations (2)
based on analytic gradient.

NOTE: A finite difference approximation is used for the derivative of
the BETA function at line

84 column 15.

WARNING: Your program statements cannot be executed completely.

WARNING: Your program statements cannot be executed completely.

WARNING: Your program statements cannot be executed completely.

WARNING: Your program statements cannot be executed completely.

WARNING: Your program statements cannot be executed completely.

NOTE: GCONV convergence criterion satisfied.

NOTE: At least one element of the (projected) gradient is greater than
1e-3.

WARNING: In a total of 5 calls an error occurred during execution of the
program statements.

NLP attempted to recover by using a shorter step size.

NOTE: There were 10001 observations read from the data set WORK.TEST.

NOTE: The data set WORK.RES has 20 observations and 8 variables.

NOTE: The PROCEDURE NLP printed pages 22-26.

NOTE: PROCEDURE NLP used (Total process time):

real time 2.84 seconds

cpu time 2.84 seconds

=20

Everything seems to be fine with the estimation process but I don't
understand this warning(i.e., convergence criteria met, small gradient
at solution, etc.).

=20

For reference, the code looks like this:

=20

proc nlp data =3D test tech=3Dquanew phessian outest=3Dres =
maxfunc=3D4000
maxiter=3D5000 ;=20

max f0;=20

=20

parms p =3D 1 ,q =3D 1, g =3D1;=20

theta =3D (p*x);

f0 =3D
-log(beta(theta,q))+q*log(g)+(theta-1)*log(m)+(theta)*log(x)-(theta+q)*l
og(g+m*x);

run;

=20

and the first 10 obs of the input data set "test" look like this:

=20

=20

Obs m x

=20

1 56.100 1

2 77.775 4

3 64.317 27

4 136.556 9

5 72.380 5

6 87.911 9

7 99.850 4

8 59.270 66

9 141.764 213

10 212.538 20

=20

=20

Can anybody explain this?

=20

Thanks,

Andy

=20

=20

=20
 | 
Pages: 1
Prev: book: Clean Code
Next: Import data from Excel