Mining for statistical enumerated type

Patent metadata

Field	Value
Publication number	US-10073894-B2
Application number	US-201514939855-A
Country	US
Kind code	B2
Filing date	Nov 12, 2015
Priority date	Dec 14, 2012
Publication date	Sep 11, 2018
Grant date	Sep 11, 2018

Abstract

Event stream attributes are analyzed to determine whether the attributes are of a statistical enumerated type, or, in other words, whether the attributes comprise statistically fixed sets of unique values, for instance. The analysis can involve determining a magnitude of change to a set of unique attribute values. In one instance, such a determination can be performed as a function of a number, or count, of unique values. Further, event stream processing can be performed dynamically, for instance, by partitioning data into time intervals and processing the intervals incrementally.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
employing at least one processor configured to execute computer-executable instructions stored in memory to perform the following acts:
receiving a diagnostic event stream associated with a software application, wherein each event in the diagnostic event stream comprises an attribute and value of the attribute;
identifying one or more unique values of the attribute from the diagnostic event stream, wherein the one or more unique values are non-duplicative;
determining a count of the unique values;
assigning the attribute to one of a plurality of categories based on variance in the count and one or more past counts of unique values; and
presenting, on a display, the attribute based on the one of the plurality of categories assigned and
generating at least one query suggestion based on the assignment of the attribute to the one of a plurality of categories for presentation on the display, the at least on query suggestion being associated with performance of diagnostics of the software application.

2. The method of claim 1 , assigning the attribute to one of a first category corresponding to a fixed set of unique values or a second category corresponding to an unfixed set of unique values.

3. The method of claim 2 further comprising producing an alert, on the display, after the attribute is reassigned from the first category corresponding to the fixed set of unique values to the second category corresponding to the unfixed set of unique values.

4. The method of claim 2 further comprises presenting, on the display, an attribute assigned to the first category corresponding to the fixed set of unique values.

5. The method of claim 4 further comprises presenting, on the display, the attribute assigned to the first category as a query suggestion.

6. The method of claim 1 , assigning the attribute to the one of the plurality of categories based on absolute and relative variance.

7. The method of claim 1 further comprises partitioning the diagnostic event stream based on a predetermined time interval.

8. The method of claim 7 further comprises identifying unique values for the attribute within a partition.

9. The method of claim 8 further comprises recording historical data and attribute category.

10. The method of claim 9 further comprises resetting the historical data upon a change in category assignment.

11. A system, comprising:
a processor coupled to a memory, the processor configured to execute the following computer-executable components stored in the memory:
a pre-process component configured to identify a count of unique values from a diagnostic event stream attribute associated with a software application, wherein the unique values are non-duplicative;
an analysis component configured to determine if the attribute comprises a statistically fixed set of values based on variance in the count and one or more past counts of unique values; and
a post-process component configured to present, on a display, an attribute that comprises the statistically fixed set of values and to generate at least one query suggestion based on the statistically fixed set of values, the at least one query being associated with performance of diagnostics of the software application.

12. The system of claim 11 , the analysis component is further configured to determine if the attribute comprises the statistically fixed set of values as a function of a mean standard deviation of the past numbers of unique values.

13. The system of claim 11 , the analysis component is further configured to determine if the attribute comprises the statistically fixed set of values as a function of a percent computed as mean standard deviation of the past numbers of unique values divided by mean of the past numbers of unique values.

14. The system of claim 11 further comprises a partition component configured to partition a diagnostic event stream associated with the diagnostic event stream attribute into segments that correspond to a predetermined time interval.

15. The system of claim 11 further comprises an extraction component configured to extract one or more unique values to extract one or more unique values of the attribute.

16. The system of claim 11 further comprises a model management component configured to manage a historical data model comprising unique attribute values, numbers of unique attribute values, and attribute state.

17. The system of claim 16 further comprises a state determination component configured to determine current attribute state based in part on the attribute state and the numbers of unique attribute values.

18. A computer-readable storage medium having instructions stored thereon that enable at least one processor to perform a method upon execution of the instructions, the method comprising:
receiving a diagnostic event stream associated with a software application;
identifying unique values for an attribute of the diagnostic event stream;
determining a state of the attribute based on a past state, past counts of unique values per interval, and a number of unique values in a current interval, wherein state indicates whether the attribute is an enumerated type or a non-enumerated type;
presenting, on a display, the attribute based on the type;
generating at least one query suggestion based on whether the attribute is an enumerated type or a non-enumerated type, the at least one query suggestion being associated with performance of diagnostics of the software application.

19. The computer-readable storage medium of claim 18 , the method further comprises clearing historical data including the past counts after transitioning from a first state indicating the attribute is the enumerated type to a second state indicating the attribute is the non-enumerated type.

20. The computer-readable storage medium of claim 18 , the method further comprises producing an alert, on the display, after transitioning from a first state indicating the attribute is the enumerated type to a second state indicating the attribute is the non-enumerated type.

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Yang Jinlin
• Xie Haibin
• Chapman Peter

Classifications

• G06F17/30539 Primary

  Physics · mapped topic

• G06F17/30598

  Physics · mapped topic

• G06F17/30516

  Physics · mapped topic

• G06F16/24568 Primary

  Data stream processing; Continuous queries · CPC title

• G06F16/285

  Clustering or classification · CPC title