Expert Network in a Method and Apparatus for Keeping and Finding Information

An apparatus for keeping and finding information, comprising:

• a textual search engine, responsive to a user query comprising a search term, using a semantic vector to promote any of documents and sites that contain other, closely related terms that strongly correlate with said search term, wherein a semantic vector of a term comprises a global frequency of said other, closely related terms within a corpus that is used to compute said semantic vector relative to said search term;
• a processor configured for generating a social graph for said user in which connections for said user comprise hyper-dimensional relationships based on semantic vectors that link said search term with a collection of said other, closely related terms;
• said processor configured for applying a personalized semantic vector for said user to said social graph, wherein said personalized semantic vector comprises a subset of said corpus that is personal to said user, said processor determining said personalized semantic vector by taking into account only documents and sites that said user has strong engagement with, and by which said user's personal semantics are determined when related to a specific term;
• and
• an expert network applying said user's personalized semantic vector to locate experts in said user's social graph based on a user query topic:
• wherein said expert network comprises a network between individuals that describes how they are related to each other in terms of expertise;
• wherein said expert network identifies said user's intent and identifies experts for said term among said user's connections;
• and
• wherein when a connection is considered an expert for said user, documents and sites that said connection kept are recommended to said user.