Skip to content

Add support for SVDD (one-class SVM) #2807

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.

Sign up for GitHub

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

Jump to bottom
Open
don-vip opened this issue Jan 31, 2014 · 12 comments
Open

Add support for SVDD (one-class SVM) #2807

don-vip opened this issue Jan 31, 2014 · 12 comments
Labels
Enhancement module:svm Needs Decision Requires decision

Comments

@don-vip
Copy link

don-vip commented Jan 31, 2014

Support Vector Data Description (SVDD) could be a nice enhancement to OneClassSVM implementation.

A technical implementation is described in this paper:
http://www.csie.ntu.edu.tw/~cjlin/papers/svdd.pdf

Source code compatible with libsvm is available here:
http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/#libsvm_for_svdd_and_finding_the_smallest_sphere_containing_all_data
@raynald
Copy link

raynald commented Mar 27, 2014

Seems nice. I'll take it.

@MichaelAquilina
Copy link

Is work underway on this or are there any plans on releasing this? I require it for some work and will attempt to implement it myself (using LibSVM) if needs be.

@ghost
Copy link

ghost commented Aug 25, 2014

Would love to see such an implementation native to scikit-learn as well

@zhaoqidong
Copy link

I want to know how to optimize the parameters (v and s), we have no test set, how to evaluate model is best ?

@amueller
Copy link
Member

zhaoqidong: open research question ;)

@MichaelAquilina : sorry for the long turn-around. There are currently no plans. I'm not sure if patching the current libsvm we have in scikit-learn is easy, but the method seems interesting. Have you worked on this in the meantime?

@ivannz ivannz mentioned this issue Nov 19, 2016
Open
@pjpjq
Copy link

pjpjq commented Oct 17, 2018

Hmm it's been 4 years...

@albertcthomas
Copy link
Contributor

@wendingp see PR #7910 which just needs some convincing arguments to explain the advantage of SVDD over One-Class SVM, see comment

@amueller
Copy link
Member

Ok so the paper has 55 citations, is that correct?

@Sorkanius
Copy link

Any news on this?

@kemalty
Copy link

kemalty commented Sep 23, 2019 •
edited
Loading

Ok so the paper has 55 citations, is that correct?

Actually this is the original paper: https://link.springer.com/article/10.1023/B:MACH.0000008084.60811.49
http://homepage.tudelft.nl/a9p19/papers/ML_SVDD_04.pdf
Has more than 2.5k citations

@kemalty
Copy link

kemalty commented Mar 11, 2020

Any updates?

@albertcthomas
Copy link
Contributor

Please check the discussion in PR #7910

@cmarmo cmarmo added module:svm Needs Decision Requires decision labels Dec 3, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
Enhancement module:svm Needs Decision Requires decision
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
10 participants
@amueller @don-vip @cmarmo @MichaelAquilina @raynald @kemalty @zhaoqidong @albertcthomas @pjpjq @Sorkanius