| 摘要 |
PROBLEM TO BE SOLVED: To provide a spam blog extraction apparatus and method for more efficiently and effectively extracting a spam blog from unknown blogs.SOLUTION: The spam blog extraction apparatus 1 includes: a determination list updating part 12 which stores a URL of a web site to be quoted by a spam blog stored in a spam blog DB 21, and adds the URL of a web site satisfying a predetermined reference as a URL for determining a spam blog to update a spam blog determining URL list 23; a spam blog determination part 14 which receives a blog to be determined and determines whether the received blog to be determined is spam or not by machine learning by using the spam blog determining URL list 23 as feature; and a candidate list extraction part 16 which, when the blog to be determined is determined as spam, extracts the URL of a web site other than a web site to be quoted by the blog to be determined which is determined as spam, as a candidate of a spam blog determining URL list. |
