Re: Inverse of a non-square matrix

From: Ekin Akoglu <ekin_at_nyahnyahspammersnyahnyah>
Date: Wed Apr 28 2010 - 14:14:11 MDT

Yes, that was what I meant, a pseudo inverse. Thank you Dennis...

Hopefully, I can write a routine to get a pseudo inverse in NCL but it
would be better if there was a function for it.

On 04/28/2010 10:52 PM, Dennis Shea wrote:
> To my knowledge, non-square matrices can not have an inverse.
> By definition, if Y is the inverse of X
>
> X*Y = Y*X
>
> However, if X is non-square the dimension sizes will not work.
>
> ===
> I think that a "pseudo inverse" can be calculated.
> However, NCL does not have have a specialized function
> to calculate this.
>
> Sorry!
> D
>
>
>
> On 4/28/10 1:09 PM, Ekin Akoglu wrote:
>> Dear all,
>>
>> Is there a function in NCL to calculate an inverse of a non-square
>> matrix? Or "inverse_matrix" function is capable of it?
>>
>> If it is capable of calculating the inverse of a non-square matrix, it
>> does not work for me since I get the error:
>>
>> On entry to DGETRI parameter number 3 had an illegal value
>>
>> I would be glad for your suggestions.
>>
>> Regards,
>>
>> Ekin
>> _______________________________________________
>> ncl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Wed Apr 28 14:14:44 2010

This archive was generated by hypermail 2.1.8 : Thu Apr 29 2010 - 08:05:27 MDT