[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

From |
"Andrew Lynch" <aalhc5@mizzou.edu> |

To |
<statalist@hsphsun2.harvard.edu> |

Subject |
st: Using multiple equations with Optimize() |

Date |
Mon, 12 Nov 2007 08:11:25 -0600 |

Stata Users,

I have run into a bit of a dilemma with programing a Kalman filter with the optimize() procedure. I have the full documentation and have not been able to clarify where exactly the optimize() command looks when it optimizes. I am minimizing the process below. My goal is to minimize the Log Likelihood function, but while I have observed data yt and xt, a and at are unobserved and my goal is to generate them. To do so, I need Stata to rerun the do loop every time it updates one of the four parameters (b1, b2, q and h). I say all that to ask the question, "Does optimize() only look to the v= equation to optimize with the parameters or does it look to the entire evaluation and move through it in some ordered fashion (thereby honoring the do loop)?" I have been unable to get this to run, and I am unsure whether it is my dumb coding error at some point or just that optimize() won't do what I want it to do.

Any enlightenment would be greatly appreciated!

Thanks,

Andrew Lynch

PhD Student

Department of Finance

Trulaske School of Busineses

University of Missouri - Columbia

/* Optimization Procedure */

void maxlik(todo, b, xt, yt, v, g, h)

{

b1=b[1]

b2=b[2]

q=b[3]

h=b[4]

i=1

do {

a[i]=b1*at[i] + b2*xt[i]

p[i]=(b1)^2*pt[i] + q

y[i]=a[i]

f[i]=ft[i] + h

at[i+1]=a[i] + ft[i+1]*p[i]*(yt[i]-y[i])

pt[i+1]=p[i] - ft[i+1]*p[i]*p[i]

L[i]=(.5)/f[i] + (.5)*(yt[i]-y[i])*(yt[i]-y[i])/f[i]

i=i+1

} while(i<(n+1))

v=sum(L)

}

S = optimize_init()

optimize_init_evaluator(S, &maxlik())

optimize_init_evaluatortype(S, "d0")

optimize_init_params(S, (0,0,0,0))

optimize_init_which(S, "min")

optimize_init_argument(S, 1, xt)

optimize_init_argument(S, 2, yt)

b=optimize(S)

*

* For searches and help try:

* http://www.stata.com/support/faqs/res/findit.html

* http://www.stata.com/support/statalist/faq

* http://www.ats.ucla.edu/stat/stata/

**References**:**Re: st: question about coefficient testing with xtgee***From:*nicola.baldini2@unibo.it

- Prev by Date:
**Re: st: question about coefficient testing with xtgee** - Next by Date:
**Re: st: capturing full survival time with KM -sts graph-** - Previous by thread:
**Re: st: question about coefficient testing with xtgee** - Next by thread:
**st: How to reproduce a very specific text book example of logistic regression using Stata** - Index(es):

© Copyright 1996–2017 StataCorp LLC | Terms of use | Privacy | Contact us | What's new | Site index |