comp.lang.idl-pvwave archive: archive

Home » Public Forums » archive » covariance matrix

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

covariance matrix [message #91809]

Sun, 30 August 2015 14:30

siumtesfai
Messages: 62
Registered: April 2013

Member

Hello all

I have filled my missing value with NAN

data = Array[4176, 168]

ntime= 168

I am interested in calculating covariance matrix

matrix = (1/ntime-1) * (Double(data) ## Transpose(data))

I found this result

result = Array[168, 168]
Min = NaN
Max = NaN

Any help would be appreciate on how to do the covariance matrix on the good data with exclusion of missing data

Best regards

Report message to a moderator

Re: covariance matrix [message #91811 is a reply to message #91809]

Sun, 30 August 2015 16:46

Jim Pendleton
Messages: 165
Registered: November 2011

Senior Member

On Sunday, August 30, 2015 at 3:30:41 PM UTC-6, siumt...@gmail.com wrote:
> Hello all
>
> I have filled my missing value with NAN
>
> data = Array[4176, 168]
>
> ntime= 168
>
> I am interested in calculating covariance matrix
>
> matrix = (1/ntime-1) * (Double(data) ## Transpose(data))
>
>
> I found this result
>
>
>
> result = Array[168, 168]
> Min = NaN
> Max = NaN
>
>
> Any help would be appreciate on how to do the covariance matrix on the good data with exclusion of missing data
>
> Best regards

Is it adequate to eliminate NaNs in your derived values rather than filter them prior to your calculation?

data = randomu(seed, 4176, 168)
data[randomu(seed, 336)*(4176*168L)] = !values.f_nan
matrix = (1/ntime - 1)*(Double(data) ## Transpose(data))
print, min(matrix)
NaN
print, min(matrix, /nan)
-1410.1838

Report message to a moderator

Previous Topic:	NG: bring_to_front doesn't always work
Next Topic:	Writing programs to use static methods but still compile in IDL 7

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Oct 08 07:15:21 PDT 2025

Total time taken to generate the page: 0.00384 seconds