Join Nostr
2024-10-20 01:03:05 UTC
in reply to

Stephen Brooks 🦆 on Nostr: npub1c9m22…z9xjd Yeah it's some sort of "effective degrees of freedom" thing, where ...

Yeah it's some sort of "effective degrees of freedom" thing, where the low amplitude principal components count as fractions of a degree of freedom.

The standard -log_2(phase space volume) information definition goes completely off the rails here, I think... or does it? You'd have to do an infinite-dimensional integral of p ln p (entropy).