System for searching
OwnerMichael Lissackinventor
Life of this patent
Estimating the term…
Abstract
A system compares two sets of database entries to prepare a list of indexed database entries based on similarity. The system is capable of providing a hypertext linked output displayed according to similarity or other user preferences, and the hypertext links are capable of querying a search engine providing links to resources related to the hypertext linked output. The user may input a source document into the system for generating a related hypertext linked output. A process parses and indexes origin database entries and source database entries and compares some or all of the entries to create the hypertext linked output according to a weighting, such as determined by a similarity search system.
Claims20
A system for converting a source document into a search portal for use by a user, comprising:
- source database entries from one or more parsed and aggregated source documents, the one or more source documents or locations to retrieve the one or more source documents being entered by the user of the system;
- origin database entries from one or more parsed and indexed origin documents;
- a similarity search system comparing a set of the source database entries to a set of the origin database entries producing a list of related links ranked at least partially by similarities determined by comparing of the set of the source database entries and the set of the origin database entries;
- constructing a hypertext linked portal for viewing of the list of related links.
A process for performing a search query comprising the steps of:
- creating a source document relational database and a repository relational database;
- conducting a search query for a search term and subdividing the search query into chunks through a parsing and searching aggregator;
- depositing the chunks from the search query into the source document relational database;
- depositing a main document repository into the repository relational database;
- subdividing the main document repository into chunks and using a similarity search software to prepare an index of relevant chunks;
- conducting a similarity search of the index of relevant chunks with the chunks from the source document relational database;
- using the parsing and search aggregator by performing a weighted aggregation by determining which relevant chunks are most similar with the chunks from the source document relational database and within confines of a number of chunks to save results parameter,
- assigning identifiers to the chunks identified as most similar and depositing the chunks into the source document relational database.
- and collating a resulting entry of similar chunks.
A method for creating a search portal, comprising:
- generating dynamically one or more source documents including text;
- parsing the text of the source documents, creating source chunks;
- aggregating the chunks into source database entries;
- parsing one or more origin documents including text, creating origin chunks of the text;
- indexing origin chunks, creating origin database entries having indexing information;
- comparing a set of the source database entries to a set of the origin database entries, using a similarity search system;
- ranking a list of related links at least partially according to similarities between the set of the source database entries and the set of the origin database entries;
- selecting a set of database entries selected from the group consisting of:
- the set of the source database entries;
- the set of the origin database entries;
- or
- both thereof;
- submitting to one or more search engines a search string selected from the set of database entries selected in the step of selecting;
- receiving output from the one or more search engines used in the step of submitting;
- constructing a hypertext linked portal for viewing of the output;
- and
- displaying the output in the hypertext linked portal, wherein the hypertext linked portal displays a link of a uniform resource locator for one or more of the source documents and the results from the one or more search engines.
Dependencies are inferred from the claim text.