Clustering Approach on Core-based and Energy-based Vibrating

Energy Description of Data Object and Clustering

“…In Table I, the number of the belonging cluster is equal to '-1' means that this data-object is the main data-object of outlier. In the view of balance problem of quantity and scale along with overabundance of data-objects in outlier, we need to use energy changing to merger the similar data-object in clusters and outliers in order to decrease the quantity of cluster, to extend the scale of cluster, to reduce the quantity of data-objects in outlier and to make the relation between similar or relative clusters prominent [9]. Through clustering with density of CRCMD algorithm, the data-objects of outlier can be submitted as a member of relative cluster to the evaluation and the decision.…”

“…But, only those clusters with enough energy can cause effective changing, and the energy is mainly from refinement and elaboration of user's theme demands. The adjust of users will cause the energy change in certain cluster or data-object and that's why they can make reaction to acquire satisfied clustering results [9].…”

“…Therefore, based on the definition of those kinds of energy [9], it's critical to put the feature description on every cluster by the methods of some models.…”