| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To extract a phrase showing a theme of a category from previously categorized document groups and to attach hierarchical tags to a document by using the extracted phrase. <P>SOLUTION: A phrase appearing at different ratios in a text and in a title is extracted as a theme of a category. Since a category is at a higher hierarchy while a theme of the category is at a lower hierarchy among a phrase showing the theme of the category, the category of each document, and the theme of the category, a category tag score, which shows probability to which category the document belongs, and a category theme tag score, which shows whether the phrase of the theme of the document belonging to the category is proper for a target document, are found. A proper combination is extracted among combinations of these scores, and the category and the category theme phrase matching the extracted combination are attached to the document as tags hierarchically. <P>COPYRIGHT: (C)2011,JPO&INPIT</p> |
