Hi Rene,
>I have just a stupid question of principle.
>After 4 weeks, do you still believe in the numerical stability of the
>results ?
>
Each time the fit function is called, it sweeps through 635 days of
stock price data for 20 different companies. I am finding fit
parameters for a 20x20 matrix (19x19 in another case). So that is more
than 190 parameters, with more than a minute per sweep. 180,000 sweeps
required by TMinuit.
For a 3x3 company matrix, the code finishes in 20 minutes or so and yes,
the fit and errors look good and I am very excited by the potential of
this technique. My 20x20 just finished over the weekend, with 226 CPU
hours logged, and I hope to see the good results this morning.
- John
--
Dr. John Krane
jkrane@netzero.com
This archive was generated by hypermail 2b29 : Sun Jan 02 2005 - 05:50:05 MET