دانلود کتاب مطالب تکمیلی برای “یادگیری ماشین احتمالی: موضوعات پیشرفته”

PageRank is not foolproof. For example, consider the strategy adopted by JC Penney, a department store in the USA. During the Christmas season of 2010, it planted many links to its home page on 1000s of irrelevant web pages, thus increasing its ranking on Google’s search engine. Even though each of these source pages has low PageRank, there were so many of them that their effect added up. Businesses call this search engine optimization; Google calls it web spam. When Google was notified of this scam (by the New York Times), it manually downweighted JC Penney, since such behavior violates Google’s code of conduct. The result was that JC Penney dropped from rank 1 to rank 65, essentially making it disappear from view. Automatically detecting such scams relies on various techniques which are beyond the scope of this chapter.