Being the information clustering device which sorts information from the similarity vis-a-vis

摘要

PROBLEM TO BE SOLVED: To classify a large amount of information at high speed. SOLUTION: Information inputted to a rough classification part 11 is roughly mapped on a self-organization map by pieces of information of high similarity and passed to a fine classification part 12. The fine classification part 12 maps the information received from the rough classification part 11 in detail to generate a self-organization map, which is further passed to a cluster recognition part 13. The cluster recognition part 13 generates a distance map from the self-organization map, finds the borders of clusters from the distance map to generate a clustering map, and extracts cluster information from the borders of clusters generated on the clustering map.