Semantic vector in a method and apparatus for keeping and finding information
View Patent ↗A semantic vector is generated for a search term based upon a global frequency of other, closely related terms within a corpus that is used to compute the semantic vector relative to the search term. The semantic vector is used in connection with a textual search engine, responsive to a user query comprising a search term, to promote any of documents and sites within results returned to the query by the search engine that contain other, closely related terms that strongly correlate with the search term.
1. A computer implemented method for keeping and finding information, comprising:
a processor extracting terms from a search query posited by a user;
for each term, said processor constructing a semantic vector by:
looking each term up in an index of terms;
filtering each term relative to any documents or sites to which said user ascribes significance by keeping said document or site;
based on said semantic vector, said processor determining whether said term is relevant to said query; and
said processor providing a find engine for user access to information via both a find interface and a keep interface, wherein a user query posited via said find engine is satisfied independently of, but may incorporate among returned search results those search results generated by, previously or concurrently, querying an independent search engine.
2. A computer implemented method for personalizing results returned to a search query, comprising:
a processor creating a semantic vector by examining words surrounding a query term in a document in context for said query term;
wherein a semantic vector is created for every word in said document;
said processor analyzing documents to which a user ascribes significance by the act of said user keeping said document;
wherein a document kept by each user has a different semantic vector than that document has for each other user;
said processor using said semantic vector for each term to personalize said search results for said user; and
said processor providing a find engine for user access to information via both a find interface and a keep interface, wherein a user query posited via said find engine is satisfied independently of, but may incorporate among returned search results those search results generated by, previously or concurrently, querying an independent search engine.
3. The method of claim 1 , further comprising:
said processor generating a personal semantic vector for said user that comprises a subset of a corpus that is personal to said user;
said processor comparing said term's semantic vector with said user's personal semantic vector; and
based on said comparing, said processor determining whether said term is relevant.
4. The method of claim 1 , further comprising:
based on said documents or sites to which said user ascribes significance by keeping said document or site, said processor boosting one or more terms and merging resulting semantic vector scores.
5. The method of claim 1 , further comprising:
said processor executing said query on said index of terms for a plurality of different terms;
said processor scoring each of said terms by matching a semantic vector for each term with query results for any one of said terms; and
said processor merging said results per document for overall relevance scoring.
6. The method of claim 2 , further comprising:
said processor any of ranking and returning search results to said user in response to a subsequent query at least in part in accordance with documents that said user kept.
7. The method of claim 2 , further comprising:
pursuant to satisfying said query, said find interface accessing both a content search index and a personal search index and, substantially simultaneously, said keep interface accessing said personal search index;
said content search index listing any of document content, user generated content, and said semantic vectors; and
said personal search index listing any of user generated content, user explicit gestures, and user implicit gestures.
8. The method of claim 2 , wherein said user explicit gestures comprise at least keeps that are captured via said keep interface and that consist of any of one or more documents and sites from among said search results to which said user ascribes significance by affirmatively keeping said document or site; and
wherein said keeps are preserved in said personal search index as one or more persistent search objects to which metadata may be attached; and
said processor any of ranking and returning search results to said user in response to a subsequent query at least in part in accordance with said keeps.