A snippet of how it works and how we describe search paths from; user to results.
The user, maybe an undergraduate physics student, frequently types search patterns like:
'standard model fermions'. From prior searches and click thrus, the personal search gadget knows some common selections are:
'wiki.standard model.fermion' or 'MIT.standard model.fermion'
where I have added the dots to indicate a link in a semantic graph. A short hand version:
(wiki,MIT).standard model.fermion
where the ' , ' is the or function, and the compact form, expanded generate two completed search paths, found in the local nested order semantic graph database. That data base is built on WebSql or equivalent. The match yields two completed search path, and ordering is maintained, the retrieval is listed by the most frequently clicked. The two completed paths get converted into web formatted search queries, mainly URL codes.
No comments:
Post a Comment