Internal Clustering Evaluation Measure

An Internal Clustering Evaluation Measure is a clustering evaluation measure that assesses clustering quality using only cluster structure without external labels.

• AKA: Internal Validation Measure, Intrinsic Clustering Measure, Unsupervised Clustering Measure.
• Context:
  • It can typically measure cluster cohesion through intra-cluster distances.
  • It can typically evaluate cluster separation via inter-cluster distances.
  • It can typically assess cluster compactness using variance measures.
  • It can often guide optimal cluster number selection through score optimization.
  • It can often validate clustering algorithm performance by quality assessment.
  • It can range from being a Distance-Based Internal Clustering Evaluation Measure to being a Density-Based Internal Clustering Evaluation Measure, depending on its measurement approach.
  • It can range from being a Single-Criterion Internal Clustering Evaluation Measure to being a Multi-Criterion Internal Clustering Evaluation Measure, depending on its evaluation factors.
  • It can range from being a Scale-Invariant Internal Clustering Evaluation Measure to being a Scale-Dependent Internal Clustering Evaluation Measure, depending on its normalization property.
  • It can range from being a Computationally-Efficient Internal Clustering Evaluation Measure to being a Computationally-Intensive Internal Clustering Evaluation Measure, depending on its complexity level.
  • ...
• Examples:
  • Silhouette Coefficient Measure for cohesion-separation balance.
  • Calinski-Harabasz Score Measure for variance ratio assessment.
  • Davies-Bouldin Index for cluster similarity evaluation.
  • Dunn Index for cluster isolation measurement.
  • ...
• Counter-Examples:
  • External Clustering Evaluation Measure, which uses ground truth labels.
  • Supervised Classification Measure, which requires labeled data.
• See: Clustering Evaluation Measure, Silhouette Coefficient Measure, Calinski-Harabasz Score Measure, Davies-Bouldin Index, Dunn Index, Cluster Validity Index, Clustering Task, Unsupervised Learning.

---