| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To specify only document aggregates with strong link, to classify the document aggregates into a cluster, and to make a role of each document in the cluster clear. <P>SOLUTION: The relativeness between the documents and the concentricity of the documents are acquired by a means 10 for specifying the document aggregate based on a specifying condition, a means 20 for evaluating the similarity between the respective documents included in the specified document aggregate, a means 30 for extracting the relativeness between the documents, based on the evaluated similarity, and a means 40 for calculating the concentricity, based on the extracted the relativeness between the documents. A topic word, the document aggregate related to the topic word, and the role of the document, are specified by a means 50 for specifying the topic word, the document aggregate related to the topic word, and the role of the document, based on acquired relativeness between the documents and the concentricity of the acquired individual documents, and are visualized to be output. <P>COPYRIGHT: (C)2008,JPO&INPIT</p> |
