This repository contains the supporting material for the talks entitled:
"An update on the latest scikit-learn features"
It was presented at PyData Paris 2024 on September 26, 2024. The slides are available here.
You will find a collection of code associated to the demo on the slides.
You can refer to the
following example
to see a use case using the new TunedThresholdClassifierCV to optimize
the decision of a classifier based on a business metric.
The following notebook show a demo leveraging the Array-API that dispatches some computation on a GPU.
This notebook contains the code shown in the talk on the metadata routing API and some notes for the context.