Validity

As described for the earlier tasks, the wildcards {participant_id} is a placeholder.

Configuration fields

parameters.clustering.validity.internal

Output

individual/{participant_id}/internal_validity.tsv (temporary file)
individual/internal_validity.tsv

Benchmarking

benchmarks/{participant_id}.internal_validity.log
benchmarks/merge_internal_validity.log

This task uses the connectivity matrix and cluster labels for each participant to compute the requested validity metrics. The sklearn package (sklearn.metrics) is used to obtain the Silhouette and Calinski-Harabasz scores, whereas the Davies-Bouldin score is implemented in cbptools.cluster.davies_bouldin_score.

The requested validity metrics are each computed per subject using the connectivity matrix as a feature array, and the predicted labels (solutions) for the participant. Note that for the Silhouette score, the metric for calculating distance between instances in the feature array is Euclidean.

Once this task is completed for each participant, the resulting scores are merged into a tab-delimited file.