| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To determine interrelated data together with the depth of the relationship, and to extract a desired number of the other data related to the data for investigation sequentially from the deeper ones of the relationship with no dropping. <P>SOLUTION: A creek extraction unit 121 extracts a maximum creek based on the related data indicating a plurality of nodes and the other node to which the plurality of nodes are related, respectively, and generates a dummy creek from the combination of the maximum creeks with the duplication rate of nodes above a predetermined value. A matrix creation unit 122 correlates the nodes and the dummy creeks with columns and rows, and creates a matrix setting 1 to a corresponding element and 0 to another element when the nodes are contained in the dummy creeks. A correspondence analysis unit 123 performs correspondence analysis on the created matrix and computes each score of the nodes and the dummy creeks. An extraction object list creation unit 124 extracts and outputs a predetermined number of the other nodes near the target node from the results of arranging the nodes in the order of the scores. <P>COPYRIGHT: (C)2010,JPO&INPIT</p> |
