INCREMENTAL CLUSTERING CLASSIFIER AND PREDICTOR

摘要

A delta cohesiveness measure is calculated for I and each child of the best host (207). The delta cohesiveness measure indicates the degree of cohesion between I and each host child. The instance I is inserted into the classification tree (208). It is determined whether more data is to be classified (209). If more data is to be classified, the next instance is received (201) and classification continues (202-209). If no more data is to be classified, then classification ends (220) resulting in a tree structure of classified instances (i.e., data points).