Detecting and modeling temporal computer activity patterns

• US 7,974,849 B1
• Filed: 05/21/2003
• Issued: 07/05/2011
• Est. Priority Date: 08/12/2002
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Hill, Rosco, Begole, James M.
• Primary Examiner(s): NGUYEN, TAN D
• Application Number:
  US10/443,176
• Time in Patent Office: 2,967 Days
• Field of Search: 705/1.1, 705/8, 705/9, 705/10, 709/224, 714/37
• US Class Current: 705/1.1
• CPC Class Codes: G06Q 10/0631 Resource planning, allocation, distributing or scheduling for enterprises or organisations
• CPC Class Codes: G06Q 10/06311 Scheduling, planning or task assignment for a person or group
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

A method is described with which to detect and model a person'"'"'s temporal activity patterns from a record of the persons computer activity or online presence. The method is both predictive and descriptive of temporal features and is constructed with a minimal amount of beforehand knowledge. Activity related data is accumulated from a mechanism that is involved in the activity of a person. Significant inactivity features are identified within the activity data. These inactivity features are characterized so as to project the temporal activity of the person. Real-time activity of the person is then detected and inactivity periods are checked for likelihood of belonging to a previously characterized significant feature. The resulting information is formatted and made available to individuals having a need for the information.

Claims

1. A method, comprising operations of:

a) accessing a database of activity data based on presence information of an individual at a location and gathered over time;

b) for a selected weekday, computing probabilities over time that said individual is present at said location during said selected weekday;

c) based on said probabilities, determining estimated features of inactivity that correspond to said selected weekday;

d) searching said database for a plurality of feature instances corresponding to a first feature of said estimated features of inactivity;

e) computing probability distributions of characteristics of said plurality of feature instances and storing, in computer memory, said probability distributions along with an identifier of said first feature as an activity rhythm model of said first feature, wherein said characteristics of said plurality of feature instances comprise a start-time, and end-time, and a duration and wherein further said probability distributions of said activity rhythm model of said first feature comprise a start-time distribution, an end-time distribution, a duration distribution, and a frequency of occurrence metric; and

f) displaying a view in a graphical user interface in a time management application, wherein the view depends on the activity rhythm model, wherein each operation of the method is executed by a microprocessor.

2. A method as described in claim 1 wherein the view does not depend on additional input as to the characteristics from a query.

3. A method as described in claim 1 further comprising:

searching said database to identify a plurality of feature instances corresponding to a second feature of said features of inactivity; and

computing probability distributions of characteristics of said plurality of feature instances of said second feature and storing, in computer memory, said probability distributions of said second feature along with an identifier of said second feature as a model of said second feature.

4. A method as described in claim 1 wherein said d) comprises comparing activity data for dates corresponding to said selected weekday, to a definition of said first feature to determine a plurality of periods of inactivity that are similar to said first feature and wherein said plurality of feature instances are said plurality of periods of inactivity.

5. A method as described in claim 4 wherein said d) further comprises determining a period of inactivity as being similar to said first feature if it has characteristics that lie within a predetermined number of standard deviations from said definition of said first feature.

6. A method as described in claim 5 wherein said definition of said first feature comprises:

a start-time;

an end-time; and

a duration.

7. A method as described in claim 1 wherein said presence information is based on periodic interaction with a keyboard at said location.

8. A method as described in claim 1 wherein said presence information is based on periodic interaction with a cursor control device at said location.

9. A method as described in claim 1 wherein said presence information is based on periodic interaction with a telephone at said location.

10. A method as described in claim 1 wherein said presence information is based on sensing by a presence sensor.

11. A method as described in claim 1 wherein said database of activity data comprises daily activity data organized by weekday and wherein, for each date, periods of activity and inactivity are recorded over time.

12. A method as described in claim 11 wherein said b) comprises:

b1) searching said database to identify activity data for dates corresponding to said selected weekday; and

b2) compiling said probabilities based on said activity data identified in said b1).

13. A method as described in claim 12 wherein said c) comprises:

c1) defining high and low probability thresholds;

c2) determining a feature of inactivity as a period having probability data that lies between said low probability threshold and said high probability threshold and wherein a start-time of said feature is defined as a transition below said low probability and wherein an end-time of said feature is defined as a transition above said high probability; and

c3) repeating said c2) for another feature of inactivity.

14. A method as described in claim 13 wherein said c) further comprises adjusting said high and low probability thresholds to eliminate insubstantial periods of inactivity.

15. A computer-readable storage medium having stored thereon a computer program, the computer program comprising a set of instructions which when executed by a computer cause the computer to perform the operations of:

a) accessing a database of activity data based on presence information of an individual at a location, wherein said database comprises records of daily activity data organized by weekday;

b) examining records of said database corresponding to a selected weekday to compute probabilities over time that said individual is active at said location during said selected weekday;

c) based on said probabilities, determining features of inactivity that correspond to said selected weekday;

d) searching said database to identify a plurality of records that contain periods of inactivity that are similar to a first feature of said features of inactivity;

e) computing probability distributions of characteristics of said periods of inactivity identified in said d) and storing, in said memory, said probability distributions along with an identifier of said first feature as an activity rhythm model of said first feature, wherein said characteristics of said periods of inactivity comprise a start-time, and end-time, and a duration and wherein further said probability distributions of said activity rhythm model of said first feature comprise a start-time distribution, an end-time distribution, and a duration distribution; and

f) displaying a view in a graphical user interface in a time management application, wherein the view depends on the activity rhythm model.

16. A computer-readable storage medium as described in claim 15 wherein the view does not depend on additional input as to the characteristics from a query.

17. A computer-readable storage medium as described in claim 16 wherein said d) further comprises determining a period of inactivity as being similar to said first feature if it has start-time and end-time characteristics that lie within a predetermined number of standard deviations from a start-time and an end-time of said first feature.

18. A computer-readable storage medium as described in claim 16 wherein said d) further comprises determining a period of inactivity as being similar to said first feature according to a similarity metric.

19. A computer-readable storage medium as described in claim 15 wherein said presence information is based on periodic interaction with a keyboard at said location.

20. A computer-readable storage medium as described in claim 15 wherein said presence information is based on periodic interaction with a cursor control device at said location.

21. A computer-readable storage medium as described in claim 15 wherein said presence information is based on periodic interaction with a telephone at said location.

22. A computer-readable storage medium as described in claim 15 wherein each record of said database of activity data comprises activity data for a respective date and records periods of activity and inactivity over time for said respective date.

23. A computer-readable storage medium as described in claim 15 wherein said c) of said operations comprises:

c1) defining high and low probability thresholds;

c2) determining a feature of inactivity as a period having probability data that lies between said low probability threshold and said high probability threshold and wherein a start-time of said feature is defined as a transition below said low probability and wherein an end-time of said feature is defined as a transition above said high probability; and

c3) repeating said c2) for another feature of inactivity.

24. A computer-readable storage medium as described in claim 23 wherein said c) further comprises adjusting said high and low probability thresholds to eliminate insubstantial periods of inactivity.