Trustrank is a method used to separate good web pages from spam. The method was first introduced in a paper by Stanford University (Gyongyi, Garcia-Molina and Pedersen) and Yahoo! in 2004 and a patent was filed.
The basic idea is similar to the method used for Page Rank (PR) in that the page link structure information is used to rank the authority of the page. The process starts by hand selecting good pages on the site (i.e. landing pages) as the trust source pages. The trust then is propogated from these pages to linked pages which then are scored. Alternately, the converse process can be applied whereby pages can be rated as spam source pages and the link propogared in the opposite direction as a measure for bad pages 9i.e. spam).
A more detailed mathematical explanation is given in http://pagerank.suchmaschinen-doktor.de/trustrank.html.