| 摘要 |
PROBLEM TO BE SOLVED: To categorize submissions into a content-related submissions and unrelated submission correctly, quickly and with low cost by automatically labeling data, although using a teacher-provided learning.SOLUTION: A content-related submission extraction device includes: micro-blog collection means for collecting a submission by using an API provided by a micro-blog; categorization means for categorizing a title of content into a title with ambiguous meaning and a title with non-ambiguous meaning; micro-blog relativeness determination means for determining a relativeness of submissions depending upon whether a title of content has ambiguous meaning. |
