Search engine providing match and alternative answers using cummulative probability values

• US 20050203883A1
• Filed: 03/11/2004
• Published: 09/15/2005
• Est. Priority Date: 03/11/2004
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Farrett, Peter W.
• Granted Patent: US 7,689,543 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

A system and method for searching a knowledge base for a match answer and an alternative answer. The method includes the steps of: inputting a search term; beginning a search at a random location in the knowledge base to identify the match answer; determining a match answer category from the match answer; determining a look-up association based on the match answer category and a search history; plugging the look-up association into an alternative answer probability table to identify an alternative answer category; and performing a secondary search at a second random location in the knowledge base to find the alternative answer that belongs to the alternative answer category.

Claims

1. A method for searching a knowledge base having a plurality of answer objects for a match answer and an alternative answer, comprising:

inputting a search term;

beginning a search at a random location in the knowledge base to identify the match answer;

determining a match answer category from the match answer;

determining a look-up association based on the match answer category and a search history;

plugging the look-up association into an alternative answer probability table to identify an alternative answer category; and

performing a secondary search at a second random location in the knowledge base to find the alternative answer that belongs to the alternative answer category.

2. The method of claim 1, wherein the match answer category and the alternative answer category form a category answer association, and the search history comprises a table of previously determined category answer associations.

3. The method of claim 2, wherein the alternative answer probability table is determined from the search history table.

4. The method of claim 1, wherein the search term is extracted from a natural language input.

5. The method of claim 1, wherein the match answer and alternative answer are presented in a natural language format.

6. A user preference search system for searching a knowledge base to find a match answer and an alternative answer for a search term, comprising:

a search engine that performs a first search at a first location in the knowledge base and returns a match answer, and performs a second search at a second location in the knowledge base to find an alternative answer, wherein the alternative answer belongs to an alternative answer category determined by plugging a look-up association into an alternative answer probability table; and

a table update system that updates the alternative answer probability table based on a table of previously determined category answer associations.

7. The user preference search system of claim 6, wherein the first and second locations are determined randomly.

8. The user preference search system of claim 6, wherein the look-up association is determined from a search history.

9. The user preference search system of claim 7, wherein each previously determined category answer association comprises a match answer category and an alternative answer category.

10. The user preference search system of claim 6, further comprising a natural language parser for receiving natural language commands and generating the search term.

11. A program product stored on a recordable medium for searching a knowledge base for a match answer and an alternative answer, comprising:

means for inputting a search term;

means for beginning a search at a random location in the knowledge base to identify the match answer;

means for selecting a match answer category from the match answer;

means for determining a look-up association based on the match answer category and a search history;

means for plugging the look-up association into an alternative answer probability table to identify an alternative answer category; and

means for performing a secondary search at a second random location in the knowledge base to find the alternative answer that belongs to the alternative answer category.

12. The program product of claim 11, wherein the match answer category and the alternative answer category form a category answer association, and the search history comprises a table of previously determined category answer associations.

13. The program product of claim 11, wherein the alternative answer probability table is determined from the search history table.

14. The program product of claim 11, wherein the search term is extracted from a natural language input.

15. The program product of claim 11, wherein the match answer and alternative answer are presented in a natural language format.