Document search method with interactively employed distance graphics display
AssigneeProximate Technologies, LLC
InventorsDumont M. Jones, Vadim M. Koganov
Life of this patent
Estimating the term…
Abstract
Method for carrying out the text content search of a document database utilizing multi-node nets. The nodes of these nets are associated with interactions and are loaded with initially selected criteria, whereupon a potential or distance function calculation is carried out with respect to each document of the database and those criteria containing nodes. The created nets are displayed with document symbols geometrically located with respect to the nodes and interaction lines in correspondence with the computed potentials to provide for an interactive visual evaluation of the search by the user. Such evaluations then may be employed to refine node contained criteria toward useful rule-based nodes.
Claims24
A method for evaluating the text content of a document database with respect to a document population, comprising the steps of:
- (a) providing a computer system having a user interface with a display;
- (b) gathering documents from said database into said system;
- (c) normalizing said gathered documents;
- (d) fingerprinting said gathered documents;
- (e) determining a text criteria with respect to said document population;
- (f) forming a net comprising at least two nodes associated by at least one interaction and displayable at said display as two or more spaced apart nodes connected by an interaction;
- (g) loading said text criteria into at least one of said nodes;
- (h) for each document of said database, calculating its geometric relative distance from a said node to derive one or more node attractors;
- (i) displaying said net at said display in combination with one or more document symbols each representing a said document located in correspondence with said calculated relative distance;
- (j) visually examining the display of said net and document symbols;
- and
- (k) determining from said document symbol locations at said display those documents, if any, which are more likely to correspond with said text criteria.
A method for evaluating the text content of a document database with respect to a population of documents comprising the steps of:
- (a) providing a computer system having a user interface with a display;
- (b) forming one or more nets each comprising at least two nodes associated by at least one interaction, one or more said nodes representing an evaluation criteria, said one or more nets being viewable at said display;
- (c) treating said documents to have an attribute value and calculating for each document a geometric relative distance with respect to a said node criteria and displaying corresponding document symbols at said display;
- (d) delimiting at said display a first region of said document symbols;
- (e) delimiting at said display a second region of said document symbols;
- (f) selecting a said document attribute value to be correlated and the criteria for establishing an attribute value match;
- (g) determining the presence of one or more document attribute value match pairs as correlations between said first and second regions;
- and
- (h) displaying said correlations at said display.
A method for searching the text content of a document database with respect to a population of documents, comprising the steps of:
- (a) providing a computer system having a user interface with a display;
- (b) identifying the population of documents to be searched;
- (c) normalizing the documents of the identified population with the steps comprising
- (c1) selecting character sequences that will separate words,
- (c2) determining to either retain or eliminate punctuation characters,
- (c3) setting regular expressions that will characterize numbers,
- (c4) setting case behavior,
- (c5) setting an offset and factor for numeric class,
- (c6) converting a document of said identified population to a character sequence,
- (c7) accessing the words, or punctuation characters, W of said character sequences,
- (c8) for each accessed W which is a number, converting such number into a sequence of word numbers, WN, and normalizing said word numbers for fingerprinting,
- (c9) marking the position and length of each W or normalized word number WN,
- (c10) for each W or normalized WN, completing said normalization by reiterating steps (c8) and (c9);
- (d) fingerprinting said normalized documents;
- (e) forming one or more nets, each comprising at least two nodes, one or more said nodes representing an evaluation criteria, said one or more nets exhibiting two or more spaced apart nodes connected by one or more interactions;
- (f) for each normalized document, calculating its geometric relative distance from a said node;
- (g) displaying said one or more nets at said display in combination with one or more document symbols representing a said document located in correspondence with said calculated relative distance;
- and
- (h) determining from said document symbol locations at said display, if any, those documents which are more likely to correspond with said evaluation criteria.
Dependencies are inferred from the claim text.