Yahoo is trying to patent the concept of "interest"

    The rating of interest on the Flickr site is some secret algorithms that in a special way dynamically calculate the absolute rating of photos on the site. To calculate the “interestingness” of a photograph, various parameters are taken into account, quantitative (number of views and comments) and qualitative: where users come from to view the photo, who comments when and who photographs, who puts it in favorites, semantic tags for this image and many other parameters.

    The algorithm for calculating interest can be compared with the well-known Google PageRank algorithm for calculating the "importance" of a web page. The difference is that Flickr’s interest takes into account many internal factors specific to social networks. In other words, the Interestingness Rating (IR) is a PR for a new generation of social search engines.

    Nobody knows the exact formula for calculating IR, and the exact value of "interest" is not shown on the screen. One can only guess how this whole system works. It is only known that the IR algorithms change periodically , just as the IR of each photo changes every day. A random set of the most interesting photos on the Flickr website over the past seven days can be viewed on one page .

    The concept of interest was first implemented on Flickr in August 2005 , along with tag clustering technology . After some time, it became clear why this technology is so important. In fact, the Flickr rating of interest is, no less, an advanced engine of recommendations for the collective mind , taking into account, moreover, the well-known phenomenon of a “long tail” (long tail).

    Obviously, the importance of IR algorithms was also appreciated by Yahoo, which filed two patents. These applications were filed February 8, 2006, and published last week.

    First patent “Ranking of media objects by their interest”reports that the rating of interest can be calculated individually for each individual user. The second patent is called Media Object Metadata Associations and Ranking . It explains what metadata can be used for ranking (tags, comments, annotations, the fact of being added to favorites). A small analysis of the text of patents can be found here .

    It should be added that some experts call Yahoo’s ridiculous attempts to patent social ranking algorithms, because these algorithms are already used in one form or another on many Web 2.0 sites.

    Also popular now: