Matrix profile as statistical distance #963
Answered
by
seanlaw
ChrisCalvin
asked this question in
Q&A
-
|
Can a matrix profile be used to compare the distance between two probability distributions? I’m wondering if the MP can be used in a Bayesian model to replace a KL Divergence. |
Beta Was this translation helpful? Give feedback.
Answered by
seanlaw
Mar 14, 2024
Replies: 1 comment
-
|
@ChrisCalvin Thank you for your question. At the end of the day, you'd essentially be computing the (z-normalized) Euclidean distance between the two distributions. I'm not sure what that would mean and you don't really need STUMPY for that as you likely wouldn't benefit from the sliding window component. |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
ChrisCalvin
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
@ChrisCalvin Thank you for your question. At the end of the day, you'd essentially be computing the (z-normalized) Euclidean distance between the two distributions. I'm not sure what that would mean and you don't really need STUMPY for that as you likely wouldn't benefit from the sliding window component.