A Better Search Engine for News
Traditional web searches are not well suited for the news. Backlink ranking methods, such as Google's PageRank, are only good in retrospect, after an article or post has been around for long enough to collect inbound links. News needs to be fresh. Standard document similarity measures do not factor in the relative significance of an article on Iraq from The Guardian compared to one from an obscure blogger.
To solve these issues, we have developed SARE (Semantic Authority Ranking Engine). It semantically analyzes all of our articles and blogs and their links, and quantitatively describes for each source their domain of authority. When you perform a search, the domain of authority for each source is consulted and used to rank the results. For example, if you
search for "iPod", you will tend to get sources like AppleInsider, Engadget, and Gizmodo.
If you
search for John McCain, you will tend to get sources like The New York Times, Politico, and Time.