search

neurosatya / AdaRGC

Adaptive riemannian geometry toolbox
GNU Affero General Public License v3.0
10 stars 3 forks source link

The value of A in function riemann_mean turns out to be all NaN #1

Open moonj94 opened 3 years ago

moonj94 commented 3 years ago

I am running the code with my own data and when I do, the while loop in rieman_mean flags an error after a few iterations because the variable A is all NaN values. Which is perplexing because nargin=1and so A = mean(B,3), and B there are no NaN values in B.

Can you please help troubleshoot this issue?

Screen Shot 2020-12-08 at 4 37 16 PM

neurosatya commented 3 years ago

Hello Jaewoong,

Thanks for your email!

How many dimensions do you have i.e. number of EEG channels? if you have a lot of channels we tend to have NaNs in covariance which will propagate to this step. Try to use shrinkage based ledoit and wolf estimator for covariance estimation.

Let me know

Best Satyam

Best regards, Satyam Kumar Doctoral student, Graduate research assistant The University of Texas at Austin satyamjuve@gmail.com | satyam.kumar@utexas.edu satyam.kumar@utexas.edu Web: neurosatya.github.io "We can only see a short distance ahead, but we can see plenty there that needs to be done." ― Alan Turing

On Wed, 9 Dec 2020 at 04:08, Jaewoong Moon notifications@github.com wrote:

I am running the code with my own data and when I do, the while loop in rieman_mean flags an error after a few iterations because the variable A is all NaN values. Which is perplexing because nargin=1 and so A = mean(B,3), and B there are no NaN values in B.

Can you please help troubleshoot this issue?

[image: Screen Shot 2020-12-08 at 4 37 16 PM] https://user-images.githubusercontent.com/53019885/101549724-c095d800-3973-11eb-8a9e-23e8aed24f92.png

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/neurosatya/AdaRGC/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHJGYENRNOW4JDSYOHSWZH3ST2TGTANCNFSM4USUCFYA .