System for assembling large databases through information extracted from text sources

• US 6,052,693 A
• Filed: 07/02/1996
• Issued: 04/18/2000
• Est. Priority Date: 07/02/1996
• Status: Expired due to Fees

• Current Assignee: Delve Software Limited
• Original Assignee: Harlequin (UK) Ltd. (Global Graphics Plc)
• Inventors: Smith, Michael J., Wilson, Adrian J., May, Alan C., Hayward-Shott, Neale, Wright, Barton D.
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): RONES, CHARLES
• Application Number:
  US08/674,502
• Time in Patent Office: 1,386 Days
• Field of Search: 705/2, 707/5, 707/3, 707/1, 707/10, 707/102, 707/104, 395/708, 704/1, 704/4
• US Class Current: 1/1
• CPC Class Codes: G06F 16/86 Mapping to a database
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing
• CPC Class Codes: Y10S 707/99948 Application of database or data structure, e.g. distributed, multimedia, or image

Abstract

Traditional information extraction processes are usually implemented on a programmed general purpose computer. The process looks for certain information, and organizes the information into a database record. The database created is usually stored in a searchable format such as a structured relational database or an object-orientated structured database, which can be accessed, research, and analyzed by computer-implemented database research systems. However, generic information extraction processes only input the extracted information into the database, in the last step of the process and do not address the problem of compiling large and comprehensive database from a plurality of source documents. Furthermore, information extraction processes are not focused on how the information extracted will be used in the construction of a large database. It would be desirable to have an information extraction system with the ability to assemble extracted information and to recognize any conflicts between the extracted information and the contents of an existing database. Accordingly, the invention is an information indexing process with the above features having the ability to construct a database with a high degree of integrity from a plurality of source documents.

Claims

1. A method for combining types of items of information from a plurality of text-based information sources in a plurality of formats, into items in a database, with

at least one information source including unstructured written text and structured text, and

at least one item of information including one or more attributes, with each attribute having at least one value,

the method comprising the steps of:

for each information source,

extracting and organizing items of information from the structured and unstructured written text of the information source, to generate an index plan, with each item in the plan organized as a hierarchic data structure representing the item'"'"'s attributes, their values and the locations of the text in the information source supporting those values; and

researching and consolidating the items of information in the index plan into items in the database to avoid duplication of items in the database.

2. The method as recited in claim 1 wherein the items can be edited by a user.

3. The method as recited in claim 1 wherein each information source can include zero or more items of information.

4. The method as recited in claim 1 wherein each information source can include zero or more types of items of information.

5. The method as recited in claim 1 wherein the step of researching includes the step of automatically formulating and applying a SOL query.

6. The method as recited in claim 1 wherein the step of researching includes the step of creating an item in the database if that item in the index plan is not in the database.

7. The method as recited in claim 1 wherein the locations of the text in the information source supporting values in the items of the index plan are preserved in the database.

8. The method as recited in claim 1 wherein each item in the database can be traced to at least one item in an index plan.

9. The method as recited in claim 1 further comprising the step of validating the integrity of the information in the items.

10. A system for combining types of items of information from a plurality of text-based information sources in a plurality of formats, into items in a database, with

at least one information source including unstructured written text and structured text, and

at least one item of information including one or more attributes, with each attribute having at least one value,

the system comprising:

an extractor configured to extract and organize items of information from the structured and unstructured written text of each information source, to generate an index plan, with each item in the plan organized as a hierarchic data structure representing the item'"'"'s attributes, their values and the locations of the text in the information source supporting those values; and

a research mechanism configured to research and consolidate the items of information in the index plan into items in the database to avoid duplication of items in the database.

11. The system as recited in claim 10 further comprising an input mechanism to allow a user to edit the items.

12. The system as recited in claim 10 wherein the research mechanism automatically formulates and applies a SOL query.

13. The system as recited in claim 10 wherein the research mechanism creates an item in the database if that item in the index plan is not in the database.

14. The system as recited in claim 10 wherein each item in the database can be traced to at least one item in an index plan.

15. The system as recited in claim 10 wherein the research mechanism determines the degree an item in the index plan matches an item in the database.

16. The system as recited in claim 10 further comprising a validator configured to validate the integrity of the information in the items.