DIVISIVE CLUSTERING DEVICE AND DIVISION DATA COUNT DECISION METHOD

摘要

PROBLEM TO BE SOLVED: To realize divisive clustering considering features of raw data. SOLUTION: A division data count decision processing part 14 clusters (sample clustering) sample data that are part of a set of data stored in a document DB 11. The division data count decision processing part 14 also evaluates the sample clustering results and decides a division data count from the evaluation results. A data division part 15 divides the set of data into the division data count decided by the division data count decision processing part 14. A clustering execution part 16 clusters each set of data divided by the data division part 15. COPYRIGHT: (C)2006,JPO&NCIPI