Item Distribution Performance
Synopsis
This operator is used for performance evaluation of flat clustering methods. It evaluates a cluster model based on the distribution of examples.
Description
The clustering operators like the K-Means and K-Medoids produce a flat cluster model and a clustered set. The cluster model has information regarding the clustering performed. It tells which examples are parts of which cluster. The Item Distribution Performance operator takes this cluster model as input and evaluates the performance of the model based on the distribution of examples i.e. how well the examples are distributed over the clusters. Two distribution measures are supported: Sum of Squares and Gini Coefficient. These distribution measures are explained in the parameters. Flat clustering creates a flat set of clusters without any explicit structure that would relate clusters to each other. Hierarchical clustering, on the other hand, creates a hierarchy of clusters. This operator can only be applied on models produced by operators that produce flat cluster models e.g. K-Means or K-Medoids operators. It cannot be applied on models created by the operators that produce a hierarchy of clusters e.g. the Agglomerative Clustering operator.
Clustering is concerned with grouping together objects that are similar to each other and dissimilar to the objects belonging to other clusters. It is a technique for extracting information from unlabeled data and can be very useful in many different scenarios e.g. in a marketing application we may be interested in finding clusters of customers with similar buying behavior.
Input
cluster model
This input port expects a flat cluster model. It is output of the K-Medoids operator in the attached Example Process. The cluster model has information regarding the clustering performed. It tells which examples are part of which cluster.
performance vector
This input port expects a Performance Vector.
Output
cluster model
The cluster model that was given as input is passed without changing to the output through this port. This is usually used to reuse the same cluster model in further operators or to view it in the Results Workspace.
performance vector
The performance of the cluster model is evaluated and the resultant Performance Vector is delivered through this port. It is a list of performance criteria values.
Parameters
Measure
This parameter specifies the item distribution measure to apply. It has two options:
- sumofsquares: If this option is selected, the sum of squares is used as the item distribution measure.
- ginicoefficient: The Gini coefficient (also known as the Gini index or Gini ratio) is a measure of statistical dispersion. It measures the inequality among values of a frequency distribution. A low Gini coefficient indicates a more equal distribution, with 0 corresponding to complete equality, while higher Gini coefficients indicate a more unequal distribution, with 1 corresponding to complete inequality.