[Thread Prev][Thread Next][Index]

Re: [ferret_users] 21 yr running correlation



I do agree with saurabh.  I will try in other software.
If we know how to do it in ferret, then that will be awesome.

On Tue, Aug 28, 2018 at 12:30 PM saurabh rathore <rohitsrb2020@xxxxxxxxx> wrote:
Dear Ryo
I want to highlight what Sandeep is asking. He is asking to plot the time series of correlation which is defined for the window of 21 years. it is similar like running trend. So it will show the variation of correlation in the window of 21 year for the time period used in the analysis.

I also searched a lot for running trend in ferret but I found its very cumbersome in ferret. I hope Sandeep will try to do in matlab like software where loop is easily applicable. This is best to my knowledge but any constructive criticism is always welcome. More over if someone is having some solution for running trend and correlation please put up in ferret  mail.

Cheers, Saurabh

On Tue, Aug 28, 2018 at 4:53 PM Ryo Furue <furue@xxxxxxxxxx> wrote:

can anyone tell me how to do 21 yr running correlation in ferret between two data sets.

If I correctly understand what "running correlation" is, I think you can use @SBX transformation to calculate the average over a fixed-width window which moves.  Then you use the ordinary definition of the correlation coefficient: mean[ (A - mean(A)) * (B - mean(B)) ] / (stddev(A) * stddev(B)), where "mean" means the box average @SBX and stddev() is also calculated over the box using @SBX.

Hope this helps,
Ryo


--


REGARDS

Saurabh Rathore
Research Scholar (PhD.)
Centre For Oceans, Rivers, Atmosphere & Land Science Technology
Indian Institute Of Technology, Kharagpur
contact :- 91- 8345984434

[Thread Prev][Thread Next][Index]
Contact Us
Dept of Commerce / NOAA / OAR / PMEL / Ferret

Privacy Policy | Disclaimer | Accessibility Statement