AN ASSOCIATIVE CLASSIFICATION METHOD FOR DETECTING USEFUL KNOWLEDGE FROM HUGE MULTI-ATTRIBUTES DATASET

摘要

PURPOSE: An association classifying method for meaningful knowledge discovery in a high-capacity multi-attribute data set is provided to predict the class level of a test set more exactly. CONSTITUTION: Data is normalized(S110). Association rules are searched in consideration of the attribute of the normalized data(S120). A classification standard value is created(S130). The classification standard value compares all rules created by the classification degree of data and the association rule search. A rule standard value is created(S140). The rule standard value decides the range of rules in the creation of the rule using the median of a class. Rules are created on the basis of a target class label(S150).