| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To determine as a spam an information source that sends spam articles designed to be ranked high in search results of a search engine when a certain keyword is used for search. Ž<P>SOLUTION: For each one of sent documents input, the document's structural typicality scores determined from the number of links and the number of paragraphs, content typicality scores determined from the ratio of the maximum number of references of links in the document by sender (or organization), and time typicality scores, i.e, the deviation in time of the document, are compared with respective preset threshold values. If the document is regarded as being created in a more typical manner than the predetermined values, the sender of the document is determined as a spam. Ž<P>COPYRIGHT: (C)2010,JPO&INPIT Ž</p> |
