Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Interest in adding online Entropy estimation? #165

Open
robertfeldt opened this issue Nov 23, 2019 · 1 comment
Open

Interest in adding online Entropy estimation? #165

robertfeldt opened this issue Nov 23, 2019 · 1 comment

Comments

@robertfeldt
Copy link
Contributor

Thanks for this great package.

I plan to implement some online Entropy estimation algorithms. Was planning to do as separate package but then realized it might fit in this one. Would this be of interest or too niche?

One technical concern might be that the calculated statistics might not be immediately available since the estimates need a certain number of data samples (in the stream) before they can be meaningfully calculated. I guess there might already be ways of handling this in the package though?!

@robertfeldt robertfeldt changed the title Interesting in adding online Entropy estimation? Interest in adding online Entropy estimation? Nov 23, 2019
@joshday
Copy link
Owner

joshday commented Nov 23, 2019

Thanks!

Your stuff could either go here or in a new package that depends on OnlineStatsBase (if you want to use the OnlineStats interface). I'm happy to have it live in OnlineStats though, under the conditions that 1) there aren't new dependencies, and 2) you have tests for your stuff. There's pros and cons of contributing vs. writing your own package so you just choose whatever is best for you.

One technical concern might be that the calculated statistics might not be immediately available since the estimates need a certain number of data samples (in the stream) before they can be meaningfully calculated

The OnlineStats interface helps with this:

  • fit! is for updating the "sufficient statistics" and doesn't necessarily update the estimate.

  • value calculates the estimate from the "sufficient statistics". Your value function could spit out error bounds/warning about low samples until you've seen a sufficient amount.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants