| Re: polynomial fitting constrained to be non-decreasing [message #81183] |
Tue, 21 August 2012 07:07  |
Russell Ryan
Messages: 122
Registered: May 2012
|
Senior Member |
|
|
On Monday, August 20, 2012 3:57:15 PM UTC-4, nata wrote:
> Hi All,
>
>
>
> I need an algorithm for polynomial fitting constrained to be non-decreasing. The idea is to have a polynomial expression fitting a set of points where its derivative is always positive.
>
>
>
> I found some information on the web but I would like to know if you have some code or ideas of how to implement that easily.
>
> Thank you for your help,
>
>
>
> nata
Hi N,
Have a look at constrained_min.pro distributed with IDL. The example they give is fitting a 4th order polynomial with constraints. It sounds like you've only got one constraint, and that the first derivative is ge 0?
-Russell
|
|
|
|
| Re: polynomial fitting constrained to be non-decreasing [message #84692 is a reply to message #81183] |
Tue, 21 August 2012 07:43  |
natha
Messages: 482
Registered: October 2007
|
Senior Member |
|
|
I don't think that the Generalized Reduced Gradient Method will solve the problem correctly...
It seems to me that a least-square polynomial "criterion" is required in order to get a good fit. Otherwise, the constraints will be respected but the final solution won't be the best one.
The best algorithm I have found is explained here:
http://www.jstor.org/stable/2984333
My poor mathematical background is not helping much... Anyway, any help will be appreciated,
nata
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
