Re: using warmabs model in ISIS

Dear Oliwia,

increasing 'relstep' or 'step' should be the correct solution for your 
problem. If I am not wrong the warmabs model uses an internal caching. 
When a parameter is only marginally varied, i.e. the change is not large 
enough (I don't know the exact treshhold), the same model configuration 
is returned. Using the standard step sizes during fitting does not 
change the model. For that reason the a gradient minimization algorithm 
assumes to have already found a minimum.

When using warmabs I also use absolute steps instead of relative steps 
(set relstep to 0 and step to the desired value) for the logarithmic 
parameters such as 'rlogxi'.

I hope that helps!
Best wishes,
Moritz

On 06/06/2013 09:34 AM, Oliwia Madej wrote:
>
> Dear ISIS users,
>
> Has any of you used warmabs model before in ISIS ? If so perhaps you encountered a problem which I'm having
> at the moment. After setting up all the variables necessary (warmabs_data, warmabs_pop) I load the model, try to fit it but the final values of the
> parameters in the model (e.g. nh,xi,vturb) don't seem to change with respect to the initial values I put in
> (which are already somewhat close to the good fit but not there yet). The parameters I mention are free of course during the fit.
> I'm more familiar with xspec and the same warmabs model is working without problems there.
> I tried playing around with relstep in set_par. It seems to make a difference if I increase the relstep to
> 0.1 for example though I'm actually not sure if this is a good solution. Perhaps some of you
> encountered this problem before and know how to sove it ?  I'd be very grateful !
>
> Cheers,
> Oliwia Madej
>
>
> ------------------------------------------------
> Oliwia Madej
> Netherlands Institute for Space Research,
> Sorbonnelaan 2, 3584 CA Utrecht
> +31 (0)88 777 5839
> ------------------------------------------------
>
>
> ----
> You received this message because you are
> subscribed to the isis-users list.
> To unsubscribe, send a message to
> isis-users-request_at_email.domain.hidden> with the first line of the message as:
> unsubscribe
>
----
You received this message because you are
subscribed to the isis-users list.
To unsubscribe, send a message to
isis-users-request_at_email.domain.hiddenwith the first line of the message as:
unsubscribe