INTERACTIVE DATA MINING

• US 20140095542A1
• Filed: 05/31/2013
• Published: 04/03/2014
• Est. Priority Date: 10/02/2012
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: ZELEVINSKY, Vladimir
• Granted Patent: US 9,201,934 B2
• Time in Patent Office: Days
• US Class Current: 707/776
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06F 16/26 Visual data mining; Browsing structured data

Abstract

A data mining system receives a data set that includes a plurality of columns of data. The system determines correlations between columns of data of the data set and displays an interactive listing of a plurality of pairs of columns based on the correlations. The listing includes preview information based on the correlations for each pair. The system receives a selection of a value from the interactive listing from a user and refines the data set in response to the selection.

Claims

1. A computer readable medium having instructions stored thereon that, when executed by a processor, cause the processor to perform data mining, the data mining comprising:

receiving a data set, the data set comprising a plurality of columns of data;

determining one or more correlations between columns of data of the data set;

displaying an interactive listing of a plurality of pairs of columns based on the correlations, wherein the listing comprises preview information for each pair, the preview information based on the correlations;

receiving a selection of a value from the interactive listing from a user; and

refining the data set in response to the selection.

2. The computer readable medium of claim 1, further comprising receiving a selection of a column of interest from the user before determining the one or more correlations, wherein the correlations comprise a correlation between the column of interest and each of the remaining columns of data.

3. The computer readable medium of claim 1, wherein the correlations comprise a pairwise correlation between the plurality of columns of data.

4. The computer readable medium of claim 1, wherein the columns of data comprise label data columns and numerical data columns.

5. The computer readable medium of claim 4, wherein each numerical data column comprises a corresponding variable.

6. The computer readable medium of claim 5, wherein the interactive listing comprises one or more label data, and for each label data a corresponding variable of a correlated numerical data variable, and wherein the selected value is a first label data.

7. The computer readable medium of claim 1, wherein the listing comprises at least one of the highest correlated or anti-correlated pairs of columns.

8. A method of data mining, method comprising:

receiving a data set, the data set comprising a plurality of columns of data;

determining one or more correlations between columns of data of the data set;

displaying an interactive listing of a plurality of pairs of columns based on the correlations, wherein the listing comprises preview information for each pair, the preview information based on the correlations;

receiving a selection of a value of one of the displayed columns from the interactive listing; and

refining the data set in response to the selection.

9. The method of claim 8, further comprising receiving a selection of a column of interest before determining the one or more correlations, wherein the correlations comprise a correlation between the column of interest and each of the remaining columns of data.

10. The method of claim 8, wherein the correlations comprise a pairwise correlation between the plurality of columns of data.

11. The method of claim 8, wherein the columns of data comprise label data columns and numerical data columns.

12. The method of claim 11, wherein each numerical data column comprises a corresponding variable.

13. The method of claim 12, wherein the interactive listing comprises one or more label data, and for each label data a corresponding variable of a correlated numerical data variable, and wherein the selected value is a first label data.

14. The method of claim 8, wherein the listing comprises at least one of the highest correlated or anti-correlated pairs of columns.

15. A data mining system for a data set comprising a plurality of columns of data, the system comprising:

a correlator that determines one or more correlations between columns of data of the data set;

a display generator that displays an interactive listing of a plurality of pairs of columns based on the correlations of the correlator, wherein the listing comprises preview information for each pair, the preview information based on the correlations; and

a data set refiner that receives a selection of a value of one of the displayed columns from the interactive listing from a user and refines the data set in response to the selection.

16. The system of claim 15, wherein the correlator receives a selection of a column of interest from the user before determining the one or more correlations, wherein the correlations comprise a correlation between the column of interest and each of the remaining columns of data.

17. The system of claim 15, wherein the correlations comprise a pairwise correlation between the plurality of columns of data.

18. The system of claim 15, comprises at least one of the highest correlated or anti-correlated pairs of columns.

19. The system of claim 15, wherein the columns of data comprise label data columns and numerical data columns.

20. The system of claim 19, wherein each numerical data column comprises a corresponding variable, the interactive listing comprises one or more label data, and for each label data a corresponding variable of a correlated numerical data variable, and wherein the selected value is a first label data.