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

From |
[email protected] |

To |
[email protected] |

Subject |
st: A challenge - Loop with an iterative algorithm for each observation |

Date |
Fri, 28 Sep 2007 14:39:16 -0300 (BRT) |

Dear all, I am working on a problem that requires the calculation of four parameters using a simple iterative algorithm. These four parameters will be used for additional calculations. Part of the the do-file is shown below. The problem is the fact I have a number X of observations (X always >= 2). Despite to the fact that my implementation works really well for _N==1, I could not think about a loop to restore the scalar z to 1 (scalar z = 1) when the first run ends. As a result, this implementation only achieves correct convergence for the first observation, since the scalar z,in the second run (2nd observation), is already smaller than the epsilon value previously specified. How could I solve this problem? part of a do-file that works well for _N==1: . . . scalar w = pAB_null scalar x = ((pAB_null*pab_null)/((pAB_null*pab_null)+(pAb_null*paB_null))) gene pAB =. gene pAb=. gene paB=. gene pab=. set more off scalar z = 1 while z > 0.000000000001 { scalar pABold = w replace pAB = ((A*2)+B+D+(E*x))/((a+b+c)*(2)) replace pAb = ((B)+(C*2)+((E)*(1-x))+(F))/((a+b+c)*(2)) replace pab = (((E)*(x))+(F)+(H)+(I*2))/((a+b+c)*(2)) replace paB = 1-pAB-pAb-pab scalar x = ((pAB*pab)/((pAB*pab)+(pAb*paB))) scalar w = pAB scalar z = abs(pABold-pAB) } . . . Again, I will be grateful for your help. Sincerely yours, Tiago * * 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/

- Prev by Date:
**st: RE: P-value not saved after -corr-?** - Next by Date:
**st: RE: RE: P-value not saved after -corr-?** - Previous by thread:
**st: about vs. verinst** - Next by thread:
**st: Update to -kdens- available from SSC** - Index(es):

© Copyright 1996–2024 StataCorp LLC | Terms of use | Privacy | Contact us | What's new | Site index |