The world’s second most-used search engine, Yahoo, is announcing a move to semantic search. Traditional search engines analyze a web page’s relevance to the search terms by analyzing the quantity of links to and from the web page. However, semantic search attempts to capture the context of data on a web page to understand the actual meaning behind the data. It does so by accessing data in XML and RDF formats from semantic networks in order to make connections between the intial search query and everything on the internet related to that search query.

Here is an example of a semantic network:

So here, we can see a semantic network pertaining to the musician Yo-Yo Ma. Semantic search embraces information found in this sort of semantic network. For example, if one were to search for “Yo-Yo Ma” in Yahoo’s proposed semantic search engine, the engine would gather data all over the internet on everything closely related to the search query, “Yo-Yo Ma”. The engine would return results that include biographies, discographies, reviews of Yo-Yo Ma’s albums and performances, musicians that Yo-Yo Ma has collaborated with in the past, sites to purchase tickets for Yo-Yo Ma’s upcoming performances, etc. This search engine system is predicted to yield higher relevancy results than those of Google’s PageRank system, which uses a link-based voting game to determine the relative importance of pages. Of course, Yahoo still has many challenges to overcome, including the difficulty of discerning the semantic information behind images and videos, especially tagless ones. And there is always the problem of the never-ending internet spam that clutters and muddles search results. But all in all, semantic search promises to deliver more comprehensive and relevant search results to improve online productivity.

Referenced Articles:

http://news.bbc.co.uk/2/hi/technology/7296056.stm

http://www2003.org/cdrom/papers/refereed/p779/ess.html