External Clustering Evaluation Measure
Jump to navigation
Jump to search
An External Clustering Evaluation Measure is a clustering evaluation measure that measures clustering quality by comparing cluster assignments with ground truth labels.
- AKA: External Validation Measure, Extrinsic Clustering Measure, Supervised Clustering Measure.
- Context:
- It can typically compare predicted clusters with reference partitions.
- It can typically measure clustering agreement through label matching.
- It can typically assess clustering accuracy via confusion matrix analysis.
- It can often validate clustering algorithms using benchmark datasets.
- It can often support algorithm comparison through standardized scoring.
- It can range from being a Pair-Based External Clustering Evaluation Measure to being a Set-Based External Clustering Evaluation Measure, depending on its comparison method.
- It can range from being a Chance-Adjusted External Clustering Evaluation Measure to being a Raw External Clustering Evaluation Measure, depending on its normalization approach.
- It can range from being a Symmetric External Clustering Evaluation Measure to being an Asymmetric External Clustering Evaluation Measure, depending on its directionality property.
- It can range from being a Binary External Clustering Evaluation Measure to being a Multi-Class External Clustering Evaluation Measure, depending on its label complexity.
- ...
- Examples:
- Counter-Examples:
- Internal Clustering Evaluation Measure, which lacks ground truth requirement.
- Unsupervised Evaluation Measure, which avoids label dependency.
- See: Clustering Evaluation Measure, Adjusted Rand Index Measure, Normalized Mutual Information Measure, Fowlkes-Mallows Index, V-Measure, Clustering Task, Ground Truth Label, Supervised Evaluation.