Several suggestions on the distance metrics of SPORF that may boost the performance on geodesic learning #346
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Current Problem
The USPORF algorithm now uses the similarity matrix as the output, but it is not good enough compared to other unsupervised machine learning methods.
Describe the solution you'd like
To make it a better algorithm for unsupervised clustering and classification, I plan to introduce different distance metrics that may help boost the classification performance.
The planned metrics are: 1) depth of nearest common ancestor and 2) length of shortest path.
The demo
The demo added implements the two new distance metrics on the two geodesic learning simulation in the USPORF paper, and both of the new distance metrics get better results than the similarity matrix on different k values and different Guassian noise dimensions.
Futhure Enhancement
I have trouble finding the infomation of tree structure and decision paths in USPORF package. If available, I believe it will boost the performance even more and if made a built-in function, it will be much convenient to use. Suggestions and instructions are welcome.
Figure showcase
Figure from the USPORF paper for reference