| Re: Regression with NaN's [message #68835] |
Wed, 25 November 2009 07:01 |
d.poreh
Messages: 406
Registered: October 2007
|
Senior Member |
|
|
On 25 Nov., 06:50, Paolo <pgri...@gmail.com> wrote:
> If you take the logarithm of a negative number, NaN is about
> as good an answer as you can hope for :)
>
> You shouldn't really throw away your data just because it's negative,
> but if you insist that you don't want those, just remove the negatives
> before calling log and you'll be all set.
>
> However, if I were the referee of your paper and you would describe
> that you just throw away negative data I would be very critical of
> that -
> might as well take away all your data and do theory instead, no? ;)
>
> Ciao,
> Paolo
>
> On Nov 25, 9:10 am, Dave_Poreh <d.po...@gmail.com> wrote:
>
>
>
>> Folks
>> I want to do some regression on my data. In my data there is an ALOG
>> function that makes my data somewhere (which I don’t know where) as
>> NaN’ss. I just want to do regression (with regress or linfit) without
>> these NaN’s.
>> Any help highly would be appreciated
>> Cheers
Good point.
Cheers
|
|
|
|
| Re: Regression with NaN's [message #68838 is a reply to message #68835] |
Wed, 25 November 2009 06:50  |
pgrigis
Messages: 436
Registered: September 2007
|
Senior Member |
|
|
If you take the logarithm of a negative number, NaN is about
as good an answer as you can hope for :)
You shouldn't really throw away your data just because it's negative,
but if you insist that you don't want those, just remove the negatives
before calling log and you'll be all set.
However, if I were the referee of your paper and you would describe
that you just throw away negative data I would be very critical of
that -
might as well take away all your data and do theory instead, no? ;)
Ciao,
Paolo
On Nov 25, 9:10 am, Dave_Poreh <d.po...@gmail.com> wrote:
> Folks
> I want to do some regression on my data. In my data there is an ALOG
> function that makes my data somewhere (which I don’t know where) as
> NaN’ss. I just want to do regression (with regress or linfit) without
> these NaN’s.
> Any help highly would be appreciated
> Cheers
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
