# Feature importance analysis

**URL:** <https://mne.discourse.group/t/feature-importance-analysis/8338>\
**Category:** Support & Discussions\
**Tags:** machine-learning\
**Created:** [February 15, 2024, 10:46am UTC](https://mne.discourse.group/t/feature-importance-analysis/8338 "2024-02-15T10:46:26Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![rodriguez-p](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/rodriguez-p/32/2130_2.png) [@rodriguez-p](https://mne.discourse.group/u/rodriguez-p)\
**Post date:** [February 15, 2024, 10:46am UTC](https://mne.discourse.group/t/feature-importance-analysis/8338/1 "2024-02-15T10:46:26Z")

</div>

Hi all,

I have succesfully run classification analyses with my EEG data using the SlidingEstimator and GeneralizingEstimator methods and a linear classifier. Now I would like to test which channels are the most informative ones for the models. I have checked some of the sklearn functions but I’m not sure if I can apply those directly to the MNE data. Anyone has done something similar and has some ideas on how to proceed?

Thanks!

---

<div class="post-metadata">

**Author:** ![agramfort](https://avatars.discourse-cdn.com/v4/letter/a/7c8e57/32.png) [@agramfort](https://mne.discourse.group/u/agramfort)\
**Post date:** [February 16, 2024, 7:58am UTC](https://mne.discourse.group/t/feature-importance-analysis/8338/2 "2024-02-16T07:58:37Z")

</div>

if you want to do feature importance analysis eg like explained in [https://inria.github.io/scikit-learn-mooc/python\_scripts/dev\_features\_importance.html](https://inria.github.io/scikit-learn-mooc/python_scripts/dev_features_importance.html)

I fear you’ll need to reimplement yourself the SlidingEstimator or hack the code in mne.

Alex