Aglomera.Evaluation.Internal Namespace

Implements the internal evaluation method in [1] that measures compactness and separation of clusters simultaneously. The numerator reflects the degree of separation in the way of how much the cluster centers are spread, and the denominator corresponds to compactness, to reflect how close the within-cluster objects are gathered around the cluster center.

Implements an internal clustering evaluation criterion as a combination (weighted average) of other IInternalEvaluationCriterion<TInstance>.

Implements the internal evaluation method in [1] that measures the "ratio of the within cluster scatter to the between cluster separation" [2].

Implements the internal evaluation method in [1] that measures the ratio between the smallest distance between observations not in the same cluster to the largest intra-cluster distance. The Dunn Index has a value between zero and infinity, and a higher index indicates a better clustering. The aim is to identify sets of clusters that are compact, with a small variance between members of the cluster, and well separated, where the means of different clusters are sufficiently far apart, as compared to the within cluster variance [2].

Implements the I-index internal evaluation method [1] that uses the ratio of the separation and compactness of a given clustering partition scheme. To measure separation, it adopts the maximum distance between cluster centers and for compactness, the distance from an to its cluster center.

Implements an internal evaluation method based on a modified/improved version of Hubert's Gamma (Γ) statistic in [1] with the transformation introduced in [2] in order to be maximized.

Implements an internal evaluation method measuring the root-mean-square standard deviation (RMSSD), i.e., the square root of the variance between all elements. This criterion considers only the compactness of the clustering partition.

Implements an internal evaluation method measuring the complement of the ratio of the sum of squared distances between elements in different clusters to the total sum of squares. This criterion considers only the separation between the clusters given some partition scheme (ClusterSet<TInstance>).

Implements an internal evaluation method that measures how similar an element is to its own cluster (cohesion) compared to other clusters (separation). The silhouette ranges from −1 to +1, where a high value indicates that the element is well matched to its own cluster and poorly matched to neighboring clusters. If most elements (average) have a high value, then the clustering configuration is appropriate. If the average is a low or negative value, then the clustering configuration may have too many or too few clusters.

Implements the within-between ratio (WB) internal evaluation method in [1] measuring the ratio of the sum-of-squares within cluster (SSW) and sum-of-squares between clusters(SSB). The result is multiplied by the negative of the number of clusters so that maximizing the ratio in some ClusteringResult<TInstance> provides the optimal partition, i.e., the optimal ClusterSet<TInstance>.

Implements the internal evaluation method in [1] known as the Xie-Beni (XB) index. It defines the inter-cluster separation as the minimum square distance between cluster centers, and the intra-cluster compactness as the mean square distance between each data object and its cluster center.

Implements the Xu-index internal evaluation method proposed in [1] measuring the compactness of clusters given some partition scheme (ClusterSet<TInstance>). The higher the negative value of the Xu-index, the better the partition in some ClusteringResult<TInstance> is.

Represents an interface for criteria which uses the internal information resulting from a AgglomerativeClusteringAlgorithm<TInstance> process to evaluate the goodness of a clustering structure without reference to external information. Implementations should be created so that when the criterion is maximized for a given ClusteringResult<TInstance> 's partition scheme, it provides the best ClusterSet<TInstance> according to that criterion.