| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To quickly and accurately determine similarity with user articles such as blogs without any burden of generating teacher data and previous learning without harmful effect. <P>SOLUTION: A blog collection part 5 acquires user articles as data from web sites via a communication network N represented by the Internet, a news acquisition part 10 acquires news as data from the web sites in a similar manner. The user articles and the news are stored in a blog storage part 25 and a news storage part 30 respectively. A quotation determination part 40 determines similarity of each of the blogs stored in the blog storage part 25 with each of news stored in the news storage part 30 by machine learning such as clustering for determining whether any of the news is quoted or not. <P>COPYRIGHT: (C)2010,JPO&INPIT</p> |
