Skip to content

Kneed algorithm fails for sharp data-sets #91

Answered by arvkevi
DoNoRe asked this question in Q&A
Discussion options

You must be logged in to vote

Hi @DoNoRe, thanks for checking out the repo. Some curves have ambiguous knee points, similar to the one from your post.

Visualizing the difference curve helps me understand why the algorithm chose the point that it did. The knee point will always be selected from a local maximum in the difference curve. In the image below I changed S=10.0 in your example data and it moves the knee much earlier.

It can be difficult to have the algorithm identify a "predetermined" knee point. I'd recommend tweaking a couple of parameters to try to get the knee point that you desire. Specifically interp_method, online, and S. You might have success fitting a line to your data, and then sending it through k…

Replies: 1 comment

Comment options

You must be logged in to vote
0 replies
Answer selected by DoNoRe
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
2 participants
Converted from issue

This discussion was converted from issue #74 on February 13, 2021 01:22.