SQL-based Naïve Bayes model building and scoring

• US 7,051,037 B1
• Filed: 05/29/2002
• Issued: 05/23/2006
• Est. Priority Date: 05/29/2002
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Thomas, Shiby, Campos, Marcos M.
• Primary Examiner(s): MIZRAHI, DIANE D
• Application Number:
  US10/156,060
• Time in Patent Office: 1,455 Days
• Field of Search: 707 1- 10, 707/104.1, 700/31, 705/26, 705/7, 711/100, 341/107, 345/440
• US Class Current: 707/737
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06F 2216/03 Data mining
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

The present invention provides an efficient method and system of data mining using SQL queries for model building and scoring. The invention provides a database management system having a database containing data, a database engine operatively connected to process the data, a SQL server operatively connected to the database and a data mining tool, whereby the data mining tool is based on a Naïve Bayes model. The SQL server uses the data and the Naïve Bayes model to develop the data mining tool. The data mining tool is located is located in the database management system. The data mining tool has a model building system based on at least one SQL query and training data, and a scoring system based on SQL queries.

Claims

1. A computer-implemented database management system, comprising:

a database containing data;

a database engine operatively connected to process said data;

a SQL server operatively connected to said database;

a data mining tool, said data mining tool based on a Naï

ve Bayes model that is built using at least a portion of said data, said SQL server using at least one SQL query to develop said data mining tool; and

wherein the Naï

ve Bayes model further comprises;

2. The computer-implemented database management system according to claim 1, wherein said data mining tool is located in said database management system, wherein said data mining tool comprises:

a model building system operable to build the Naï

ve Bayes model using at least one SQL query and training data; and

a scoring system operable to score the Naï

ve Bayes model using SQL queries.

3. The computer-implemented database management system according to claim 2, wherein said model building system and said scoring system arrange said data into a plurality of groups.

4. The computer-implemented database management system according to claim 2, wherein said scoring system comprises making predictions based on said model building system.

5. The computer-implemented database management system according to claim 1, wherein said data mining tool is located in said database management system, wherein said data mining tool comprises a model building system operable to build the Naï

ve Bayes model using at least one SQL query and training data.

6. The computer-implemented database management system according to claim 5, wherein said model building system comprises developing a set of rules from said training data.

7. The computer-implemented database management system according to claim 1, wherein said data mining tool is located in said database management system, wherein said data mining tool comprises a scoring system operable to score the Naï

ve Bayes model using SQL queries.

8. A computer-implemented database management system, comprising:

a database containing data;

a database engine operatively connected to process said data;

a SQL server operatively connected to said database;

a data mining tool, said data mining tool based on a Naï

ve Bayes model that is built using at least a portion of said data, said SQL server using at least one SQL query to develop said data mining tool; and

wherein the Naï

ve Bayes model further comprises;

P(ABC . . . |this-prediction)=P(A|this-prediction)P(B|this-prediction)P(C|this-prediction).

9. A computer-implemented method for performing data mining, comprising:

receiving data from a database;

issuing at least one SQL query to said database;

mining data using at least one SQL query, said mining based on a

Naï

ve Bayes model; and

wherein said Naï

ve Bayes model is based on;

10. The computer-implemented method according to claim 9, wherein said mining data comprises:

building a model using at least one said SQL query and a set of training data; and

scoring said model using at least one SQL query.

11. The computer-implemented method according to claim 10, wherein said building a model and said scoring a model comprises arranging said data into a plurality of groups.

12. The computer-implemented method according to claim 9, wherein said mining data comprises building a model using at least one SQL query and a set of training data.

13. The computer-implemented method according to claim 12, wherein said building a model comprises developing a set of rules from said training data.

14. The computer-implemented method according to claim 9, wherein said mining data comprises scoring said model using at least one SQL query.

15. The computer-implemented method according to claim 14, wherein said scoring a model comprises developing predictions based on said model.

16. A computer-implemented method for performing data mining, comprising:

receiving data from a database;

issuing at least one SQL query to said database;

mining data using at least one SQL query, said mining based on a Naï

ve Bayes model; and

wherein said Naï

ve Bayes model further comprises;

P(ABC . . . |this-prediction)=P(A|this-prediction)P(B|this-prediction)P(C|this-prediction).

17. A computer-readable program product for performing data mining, comprising:

a database containing data;

a database engine operatively connected to process said data;

a SQL server operatively connected to said database;

a data mining tool, said data mining tool based on a Naï

ve Bayes model that is built using at least a portion of said data, said SQL server using at least one SQL query to develop said data mining tool; and

wherein said Naï

ve Bayes model comprises;

18. The computer-readable program product according to claim 17, wherein said data mining tool is located in said database, wherein said data mining tool comprises:

a model building system operable to build the Naï

ve Bayes model using at least one SQL query and training data; and

a scoring system operable to score the Naï

ve Bayes model using SQL queries.

19. The computer-readable program product according to claim 17, wherein said data mining tool is located in said database, wherein said data mining tool comprises a model building system operable to build the Naï

ve Bayes model using at least one SQL query and training data.

20. The computer-readable program product according to claim 19, wherein said model building system comprises developing a set of rules from said training data.

21. The computer-readable program product according to claim 17, wherein said data mining tool is located in said database, wherein said data mining tool comprises a scoring system operable to score the Naï

ve Bayes model using SQL queries.

22. The computer-readable program product according to claim 21, wherein said scoring system comprises making predictions based on said model building system.

24. The computer-readable program product according to claim 18, wherein said model building system and said scoring system arrange said data into a plurality of groups.

23. A computer-readable program product for performing data mining, comprising:

a database containing data;

a database engine operatively connected to process said data;

a SQL server operatively connected to said database;

a data mining tool, said data mining tool based on a Naï

ve Bayes model that is built using at least a portion of said data, said SQL server using at least one SQL query to develop said data mining tool; and

wherein said Naï

ve Bayes model further comprises;

P(ABC . . . |this-prediction)=P(A|this-prediction)P(B|this-prediction)P(C|this-prediction).