Region Based Information Retrieval System

• US 20110238664A1
• Filed: 03/25/2011
• Published: 09/29/2011
• Est. Priority Date: 03/26/2010
• Status: Active Grant

• Current Assignee: Palle M. Pedersen
• Original Assignee: Palle M. Pedersen
• Inventors: Pedersen, Palle M.
• Granted Patent: US 8,650,195 B2
• Time in Patent Office: Days
• US Class Current: 707/737
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence
• CPC Class Codes: G06F 16/93 Document management systems

Abstract

A region based information retrieval system improves on conventional information retrieval systems by breaking down documents into one or more region(s) and processing the additional information available at a region level of analysis. When looking at regions, it becomes possible to quickly distinguish between groups of related documents, quickly ignore or focus on certain information, track recent evolutions of documents, as well as understand the historical relationships, heritage, and versions of these documents. This is all possible whether or not the document publishers specify where the content originally came from.

Claims

1. A system for retrieving information from document collections comprising:

a document collection subsystem for managing document collection in a document database;

a region finding, splitting and graphing subsystem for analyzing documents in the document database, establishing regions of these documents, identifying region sets of identical or nearly identical regions across documents, and storing these regions sets in a region set database;

an indexing subsystem for making the region sets searchable and storing the index information in a searchable index of region set database; and

a searching and ranking subsystem for finding region sets in the region set database using the searchable index of region set database based on an information request, and communicating the search results to end-users or another system.

2. The system of claim 1, wherein the region finding, splitting and graphing subsystem further establishes region set graphs with relationships between the region sets, and stores these region set graphs in a region set graph database;

3. The system of claim 1, wherein the searching and ranking subsystem further creates a list of documents from the document database which contains regions belonging to the region sets found and communicates these results to end-users or another system.

4. The system of claim 1, wherein the searching and ranking subsystem further creates a list of region set clusters of closely related region sets from the regions sets found, where the relations are based on the relationships in the region set graphs obtained from the region set graphs database and communicates these results to end-users or another system.

5. The system of claim 4, wherein the results contain the regions sets within each regions set cluster.

6. The system of claim 5, wherein the regions sets within each regions set cluster are ranked based on the chronology of the region sets.

7. The system of claim 4, wherein the results contain the documents from the document database which contains regions belonging to the region sets found.

8. The system of claim 1, wherein the information request comprises a search text.

9. The system of claim 1, wherein the information request comprises similarity to a document, image, video, or other content.

10. A method for retrieved information from document collections, comprising:

identifying documents related to an information request;

establishing any relationships between these documents;

clustering the documents based on such relationships;

ranking each cluster of documents; and

communicating the search results to end-users or another system.

11. The method of claim 10, wherein a chronology influences the ranking of clusters or the ranking of documents within a cluster.

12. The method of claim 10, wherein the search results comprise the documents in each identified cluster.

13. A method of organizing and retrieving information in documents collections comprising:

breaking down each document into one or more regions;

establishing region sets which are duplicates or near duplicates of each other;

searching and finding region sets related to an information request; and

communicating the search results to end-users or another system.

14. The method of claim 13, further comprising:

establishing region set graphs with the relationships between the region sets; and

clustering the resulting region sets into region set clusters of closely related region sets from the regions sets found, where the relations are based on the relationships in the region set graphs.

15. The method of claim 14, wherein the relationship between these region sets are either one region set being a partial or full subset of the other region set or one region set containing a reference to or keywords of a document containing all or part of the other region set.

16. The method of claim 14, wherein an attempt is made to establish the strength of the relationship between the region sets.

17. The method of claim 14, wherein an attempt is made to establish the chronology of the region sets within region set clusters.

18. The method of claim 14, wherein the search results are ranked based on the relationships between the region sets.

19. The method of claim 13, wherein the search results are the documents partially or fully containing the identified sets of regions.