Guillaume Lemaitre
Guillaume Lemaitre
Reflection around a unified API to compute feature importance in scikit-learn
Here is the SLEP regarding the Outlier Rejection API
TODO: - [ ] Integrate the heart-rate model into the `sksports.model` module. At a first glance, it would be good to not depend on `lmfit` but only `scipy` it will...
This is a ticket to design the package which should take care about the IO of cycling/sports file. Currently: * `sweatpy` allows to read data from GoldenCheetah and Strava; *...
Partially addressing #5 Implement a scikit-learn like regressor to fit and predict heart-rate data from power.
To release the first scikit-sports, we will need: - [ ] Create the multibuild as in `scikit-cycling-wheels` - [ ] Create a conda-forge feedstock - [ ] Push the wheel...
Implement a new parameter `recalibrate` to correct for sampling bias introduced due to sampling in ensemble classifiers ### TODO: - [x] implement the same trick for all ensemble classifiers -...
We should create a function which should load a trained model and produce some power estimation.
We will need to handle the IO (fit, etc.) in an outside package similarly to imageio. We would benefit to be consistent.
This is an issue to report the step to actually take over the work of @marctorsoc in https://github.com/scikit-learn/scikit-learn/pull/23183 and split the PR into smaller one to facilitate the review process....