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

From |
"Mentzakis, Emmanouil" <e.mentzakis@abdn.ac.uk> |

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

Subject |
st: RE: Re: Calculation of integrals |

Date |
Tue, 11 Dec 2007 18:24:39 -0000 |

Dear Rodrigo, Thanks for your reply. Differences in the results are not restricted to a specific integral. E.g. using the command below clear range x 1 100 50 g double lnx=ln(x) g double y=5*lnx integ y x, g(result) If I understand correctly, I compute the integral with lower limit 1 and upper limit 100, over 50 points and the function integrated is 5*ln(x). Is this correct? Similarly in Limdep, I thought that the same calculation would be done by the command Fintegrate ;fcn = 5*log(var) ;labels = var ;start = 10 ;pts =50 ;limit = 1,100 ;vary(var) $ Where, again, the limits are from 1 to 100, for 50 points and for the function 5*log(x). What is specified as "start" does not influence the calculation. It, actually, is a value given so that it can be assessed before the computation is done whether the function evaluation is possible at all. I'd really appreciate if you could comment on the equality of the two commands, because probably I am estimating two different things without realising it. Thanks Manos -----Original Message----- From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Rodrigo A. Alfaro Sent: 11 December 2007 00:47 To: statalist@hsphsun2.harvard.edu Subject: st: Re: Calculation of integrals Manos, It is difficult to give an answer with this info. What is exactly the integral that you want to calculate? What is the initial parameter in Limdep? It seems that you are trying to get an integral of the normal cdf, then something like 1000 seems too much as result (just a guess) Rodrigo. ----- Original Message ----- From: "Mentzakis, Emmanouil" <e.mentzakis@abdn.ac.uk> To: <statalist@hsphsun2.harvard.edu> Sent: Monday, December 10, 2007 4:47 PM Subject: st: Calculation of integrals > Dear all, > > I am calculating an integral in Stata and then compare the calculation > with the same calculation in Limdep. > > However, the difference between the two calculations is very big (Stata > = 3.233 ; Limdep = 980.525). > > Below I have the code for Stata and the code for Limdep. > > Could anyone explain if I am making a mistake or what is the reason for > this result? > > Thank you > Manos > > > ***************************** > *********** Stata *********** > ***************************** > > clear > set obs 300 > capture drop xb x > g xb=invnormal(uniform()) > range x .001 1000 300 > g double lnx=ln(x) > g double norm=normal(xb + .5*lnx)/300 > integ norm x, g(integral) > > number of points = 300 > > integral = 3.233067 > > > > ******************************** > ************ Limdep ************ > ******************************** > Coping the data from Stata, in Limdep we don't have to create a variable > for the range. It automatically uses the variable that is put in the > "label" and its limit is given by the "limit" and the number of points > by "pts". "Start" is the values of the parameters, but changing it > doesn't make any difference in the actual calculation. > > Fintegrate ;fcn = PHI(xb + .5*log(var))/300 > ;labels = var > ;start = 500 > ;pts =300 > ;limit = 0.001,1000 > ;vary(var) > $ > +---------------------------------------------------+ > | Function integration: | > | Grid is 300 points in [ .001 to 1000.000] | > | Value of the integral is 980.52567 | > +---------------------------------------------------+ > > > > > * > * 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/ * * 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/ * * 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/

**Follow-Ups**:**Re: st: RE: Re: Calculation of integrals***From:*"Scott Merryman" <scott.merryman@gmail.com>

- Prev by Date:
**AW: AW: st: panel-specific autocorrelation with unbalanced panels** - Next by Date:
**st: Re: SQL Query Password/User ID** - Previous by thread:
**st: Average and count from a given number of previous observations** - Next by thread:
**Re: st: RE: Re: Calculation of integrals** - Index(es):

© Copyright 1996–2021 StataCorp LLC | Terms of use | Privacy | Contact us | What's new | Site index |