Search engine providing match and alternative answer

• US 20030220909A1
• Filed: 05/22/2002
• Published: 11/27/2003
• Est. Priority Date: 05/22/2002
• Status: Active Grant

• Current Assignee: International Business Machines SA (International Business Machines Corporation)
• Original Assignee: International Business Machines SA (International Business Machines Corporation)
• Inventors: Farrett, Peter W.
• Granted Patent: US 7,107,261 B2
• Time in Patent Office: Days
• US Class Current: 707/3
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

A search engine that is client-side, provides high performance and creates a user preference history record relative to a knowledge base is provided. The more time the user spends using the search engine, the more the history record is updated, refined, and translated into personal preferences relative to that knowledge base. The invention is natural language based. In addition, it is applicable to a variety of knowledge bases, each which may have their own associated history record, and can be adapted to other applications such as on-line help, interactive training, wizard functions, virtual chat sessions, intelligent bots, etc.

Claims

1. A method of searching for data in a knowledge base having a plurality of categories based on a search keyword, each category having respective keyword data, the method comprising the steps of:

determining a match answer category by searching the keyword data for a match to the search keyword;

assigning a first weighted probability value to each category having a keyword that matches the search keyword;

assigning a second weighted probability value to each category having a shared keyword, other than the search keyword, in respective keyword data to that of the match answer category;

assigning a third weighted probability value to each category related to the match answer category in a relation table;

updating cumulative probability values of a history record with the assigned weighted probability values; and

determining an alternative answer category based on a highest cumulative probability value for the match answer category.

2. The method of claim 1, further comprising the step of loading the knowledge base into read only memory.

3. The method of claim 1, wherein when more than one category is determined to be the match answer category, a first occurrence is reported as the match answer category.

4. The method of claim 1, wherein the history record compares each category with each other category based on the cumulative probability values.

5. The method of claim 1, further comprising the step of reporting the match answer category and the alternative answer category.

6. The method of claim 1, further comprising the step of saving the history record after updating.

7. The method of claim 1, wherein the search keyword is parsed from a natural language input string.

8. The method of claim 1, wherein the step of updating the history record includes adding the assigned weighted probability values to saved cumulative probability values.

9. A search engine system for searching a knowledge base having a plurality of categories based on a search keyword, the search engine system comprising:

a match module that determines a match answer category based on a search keyword;

an alternative answer module including;

a probability value assigner that;

assigns a first weighted probability value to each category having a keyword that matches the search keyword;

assigns a second weighted probability value to each category having a shared keyword, other than the search keyword, in respective keyword data to that of the match answer category;

assigns a third weighted probability value to each category related to the match answer category in a relation table;

an updater that updates a history record of cumulative probability values created over time based on the assigned weighted probability values; and

an answer determinator that determines an alternative answer category based on a highest cumulative weighted probability value for the match answer category.

10. The search engine system of claim 9, wherein the knowledge base is loaded into read only memory on a client-side system.

11. The search engine system of claim 9, further comprising a reporter for reporting the match answer category and the alternative answer category.

12. The search engine of claim 11, wherein when more than one category is determined to be the match answer category, a first occurrence is reported as the match answer category.

13. A computer program product comprising a computer useable medium having computer readable program code embodied therein for searching a knowledge base based on a search keyword, the program product comprising:

program code configured to determine a match answer category by searching the keyword data for a match to the search keyword;

program code configured to assign a first weighted probability value to each category having a keyword that matches the search keyword;

program code configured to assign a second weighted probability value to each category having a shared keyword, other than the search keyword, in respective keyword data to that of the match answer category;

program code configured to assign a third weighted probability value to each category related to the match answer category in a relation table;

program code configured to update cumulative probability values of a history record with the assigned weighted probability values; and

program code configured to determine an alternative answer category based on a highest cumulative probability value for the match answer category.

14. The program product of claim 13, wherein when more than one category is determined to be the match answer category, a first occurrence is reported as the match answer category.

15. The program product of claim 13, wherein the history record compares each category with each other category based on the cumulative probability values.

16. The program product of claim 13, further comprising program code configured to report the match answer category and the alternative answer category.

17. The program product of claim 13, further comprising program code configured to save the history record after updating.

18. The program product of claim 13, wherein the search keyword is parsed from a natural language input string.

19. The program product of claim 13, wherein the program code configured to update the history record includes program code configured to add the assigned weighted probability values to saved cumulative probability values.

20. The program product of claim 13, wherein the computer useable medium is loaded to a client-side system.