"Estimating the dimensionality of neural data brings its own unique challenges. [D]imensionality can be defined as the number of linear orthogonal components (singular- or eigenvalues) underlying a matrix that are larger than zero (Shlens, 2014)"https://doi.org/10.1101/232454
I'm not sure but given all the noise I am sure it does give output that's a real number. But I am no expert! @ProfData can explain more, I am sure.
-
-
I did read the whole paper, with interest. Looks like really useful ideas so look forward to seeing it used more. Thanks for the link.
-
You're way ahead of me, I have not read the whole thing yet!
End of conversation
New conversation -
-
-
Thanks! Yes and no on fractional dimensionality. For SVD for each cross-validation fold, the winning model is an integer dimensionality. However, when these results combine and go into the Bayesian hierarchical model, then it becomes continuous between min and max dimensionality.
-
Thank you!
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.