| 摘要 |
PROBLEM TO BE SOLVED: To arrange a document in a space of specified dimesions at a high speed according to the perspective of a semantic content. SOLUTION: A clustering means 104 clusters the appropriate number of vectors for a document vectorized by a vectorizing means 103, and a cluster center extracting means 105 extracts a point representing each cluster, and a mapping means 106 is applied only to those cluster centers. A cluster developing means 107 decides the arrangement place of the vectors included in each cluster from the results of the clustering means 104 and the mapping means 106. |
