Clustering.md

Clustering

Datasets used as gold standard

Dataset	Interpretation of clusters	Clusters	Size
Teams	{Football Teams, Basketball Teams}	2	4,206
(Balanced) Cities and Countries	{Cities, Countries}	2	4,344
Cities, Albums, Movies, AAUP, Forbes	{Cities, Albums, Movies, Universities, Companies}	5	6,357
! Cities and Countries	{Cities, Countries}	2	11,182

Model and its configuration

Model	Configuration
Agglomerative Clustering	similarity_metric
Ward Hierarchical Clustering	similarity_metric
DBSCAN	similarity_metric
k-Means	-

For each missing entity a singleton cluster is created, i.e. a cluster which contains only the current entity. Further, soft clustering approaches, such as DBscan, do not cluster all entities. We call these entities miss-clustered entities and manage them exactly as the missing entities, i.e., we create a singleton cluster for each of them. The evaluation metrics are applied to the combination of the clusters returned by the clustering algorithm and all the singleton clusters.

Output of the evaluation

Metric	Range	Optimum
Adjusted rand score	[-1,1]	Highest
Adjusted mutual info score	[0,1]	Highest
Fowlkes Mallow index	[0,1]	Highest
v_measure score	[0,1]	Highest
Homogeneity score	[0,1]	Highest
Completeness score	[0,1]	Highest