[Scilab-users] Iterative regresion

[Dang, Christophe]

Hello, 

-----Message d'origine-----
De la part de samaelkreutz
Envoyé : mercredi 12 mars 2014 08:14

> My model has the following form:
>
> log(PGA) = d1*e1 + d2*e2 +...+d(max number of M bins)*e(max number of M bins) + a*log(R/Ref) + b*(R-Ref)
> [...]
> x=MI\log(PGA);
> [...]
> My question is:  I obtained the foliowing coeff.
> But some of them doesn't fit well!
> My teacher said "Make a simple regression, obtain coefficients
> and then iterate again to get the final coefficients"

I'm afraid I don't understand very well your problem.
Probably a lack of vocabulary on my side,
I don't know what a "M bin" is,
I don't use this in my job.

The backslash performs a linear regression
(as long as MI is not square).
What do you mean by "simple regression" ?

Additionally, you use a W and don't explain what it is.
>From your example, it is not clear which coefficients you try to refine
(the d's, the e's ?)

Maybe you could send a full (and not too complicated) script that we can run, and point out why the values you get are not satisfying.

Regards.

-- 
Christophe Dang Ngoc Chan
Mechanical calculation engineer

______________________________________________________________________

This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error), please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.
______________________________________________________________________