[Conformal EEG] K-means clustering

[lehendo]

This PR adds the k-means clustering conformal prediction.

• Implemented ClusterLabel for cluster-specific calibration thresholds using K-means on patient embeddings
• Added example script tuev_kmeans_conformal.py demonstrating usage on TUEV EEG dataset
• Includes test suite with coverage for initialization, calibration, and prediction

[Copilot]

Pull request overview

This PR adds K-means clustering-based conformal prediction (ClusterLabel) for multiclass classification in EEG analysis. The method groups patients into clusters using K-means on embeddings and computes cluster-specific calibration thresholds to improve prediction set efficiency compared to global thresholds.

Changes:

• Implemented ClusterLabel class that performs K-means clustering on patient embeddings and applies cluster-specific calibration thresholds
• Added comprehensive test suite covering initialization, calibration, and prediction
• Included example script demonstrating usage on TUEV EEG dataset with ContraWR model

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
pyhealth/calib/predictionset/cluster/cluster_label.py	Core implementation of cluster-based conformal prediction with K-means clustering
pyhealth/calib/predictionset/cluster/init.py	Module initialization exporting ClusterLabel class
pyhealth/calib/predictionset/init.py	Updated to export ClusterLabel from cluster submodule
tests/core/test_cluster_label.py	Test suite covering initialization, calibration, and forward pass with various configurations
examples/conformal_eeg/tuev_kmeans_conformal.py	Example script demonstrating ClusterLabel usage on TUEV EEG events dataset

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[lehendo]

Tbh, the copilot suggestions were pretty valid. Implemented some of them.

[siddharthal]

lgtm