 | Clusters Class |
Inheritance HierarchyNamespace: Novacta.Analytics
Assembly: Novacta.Analytics (in Novacta.Analytics.dll) Version: 2.0.0
SyntaxThe Clusters type exposes the following members.
Methods| Name | Description | |
|---|---|---|
   | Discover |
Discovers optimal clusters
in a data set.
|
| Explain |
Explains existing clusters by selecting
a number of features from the specified corresponding data set.
|
RemarksThe Clusters class supplies method Discover(DoubleMatrix, Int32) for identifying clusters in a data set, and method Explain(DoubleMatrix, IndexPartitionDouble, Int32) to select features from a data set in order to interpret an existing partition of the available items.
Such methods assume that the data set can be represented as a DoubleMatrix instance having as many rows as the number of items, while the number of columns is the number of available features. The i-th row thus contains the observations at item i of the features under study.
Clusters of items are described as an IndexPartitionT instance, whose parts represents the clusters as collections of indexes valid for the rows in the data set.
Advanced scenarios
Class Clusters exploits the Cross-Entropy method to solve its optimization problems. For more information about specialized algorithms, or different criteria for cluster discovering or feature selection, see the documentation of methods Discover(DoubleMatrix, Int32) or Explain(DoubleMatrix, IndexPartitionDouble, Int32), respectively.See Also