Re: Polyfit with scaling

[Rolf Fabian]

Ben Abbott wrote:
> 
> I don't believe Octave has any functions that are restricted to real  
> values. So ... yes, polyfit accepts and returns complex values.
> 
> Ben
> 

Lets say that way: The majority of Octave's mapping functions is
able to handle complex arguments. But unfortunately some
important functions do not and need to be improved.

E.g.:
octave-3.0.0.exe:1> x = fsolve('x.^2+1',0.9999*i)
error: fsolve: iteration is not making good progress
        not returning expected x=+/-i

octave-3.0.0.exe:2> gamma (x=1+i)
error: gamma: unable to handle complex arguments
      Handling complex x requires the Lanczos approximation
      of the gamma function which is  unfortunately
      not implemented in Octave.     

octave-3.0.0.exe:3> erf(1+i)
error: erf: unable to handle complex arguments

octave-3.0.0.exe:4> erfc(1+i)
error: erfc: unable to handle complex arguments

Form 3 of the qz decomposition

Rolf


-----
Rolf Fabian
<r dot fabian at jacobs-university dot de>

-- 
View this message in context: 
http://www.nabble.com/Polyfit-with-scaling-tp15191415p15264392.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.