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
Added SignatureVectorizer (iisignature) #108
Conversation
Codecov Report
@@ Coverage Diff @@
## master #108 +/- ##
==========================================
+ Coverage 90.85% 91.13% +0.28%
==========================================
Files 32 34 +2
Lines 4649 4774 +125
==========================================
+ Hits 4224 4351 +127
+ Misses 425 423 -2
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
This looks like a good addition. Would you be willing to add some basic tests, and possibly even a tutorial guide for the documentation? Also is it worth adding iisignature to the requirements? |
No problem, I should have some time over the next few weeks to add tests and tutorial guides. Happy to add iisignature to the requirements too - the current implementation was in case we wanted to avoid having too many dependencies, but if that's not an issue I'll add it in directly. |
Thanks -- I know the tests and documentation are a lot of extra work, but they are very valuable for maintenance, and for ensuring users can find and use the feature, so it will definitely be worth the effort in the long run. As far as dependencies go ... I haven't checked what iisignature requires. If it is reasonably self contained then one extra dependency is probably fine for now. |
…orm in SignatureVectorizer
I have added an example notebook and some tests for the SignatureVectorizer. I'll add more functionality (e.g. common path transforms) and more examples in the future too, but this should suffice as a good starting point.
|
…l to azure-pipelines.yml, before it installs requirements.txt
Thanks for this -- as I say it should be very valuable for a great many people and is a great complement to the techniques we had so far. I'm looking forward to putting this to use myself. |
I've implemented SignatureVectorizer, which returns the path signatures for a collection of paths.
This vectorizer essentially wraps the iisignature package such that it fits into the standard sklearn style fit_transform pipeline. While it does require iisignature, the imports are written such that the rest of the library can still be used if the user does not have iisignature installed.
For more details on the path signature technique, I've found this paper quite instructive:
A Primer on the Signature Method in Machine Learning (Chevyrev, I.)