Generating predilection cohorts

• US 8,117,144 B2
• Filed: 12/12/2008
• Issued: 02/14/2012
• Est. Priority Date: 12/12/2008
• Status: Active Grant

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Angell, Robert Lee, Friedlander, Robert R, Kraemer, James R
• Primary Examiner(s): Starks, Wilbert L
• Application Number:
  US12/333,323
• Publication Number:
  US 20100153353A1
• Time in Patent Office: 1,159 Days
• Field of Search: 706/45
• US Class Current: 706/45
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A computer implemented method, apparatus, and computer program product for generating risk scores for specific risk cohorts. Digital sensor data associated with a specific risk cohort is received from a set of multimodal sensors. The specific risk cohort includes a set of identified cohort members. The digital sensor data includes metadata describing attributes associated with at least one cohort member in the set of identified cohort members. Description data for each cohort member in the set of identified cohort members is retrieved to form a set of cohort description data. The description data for each cohort member comprises data describing a previous history of the cohort member or a current status of the cohort member. The cohort member is a person, animal, plant, thing, or location. A specific risk score is generated for the specific risk cohort based on selected risk factors, the attributes associated with the at least one identified member, and the set of cohort description data. A response action is initiated in response to a determination that the specific risk score exceeds a risk threshold.

Claims

1. A computer implemented method of generating a predilection cohort, the computer implemented method comprising:

receiving digital sensor data associated with a predilection cohort from a set of multimodal sensors, wherein the predilection cohort comprises an identified member of the predilection cohort, and wherein the digital sensor data comprises metadata describing attributes of the identified member;

processing and parsing the digital sensor data using a set of data models to identify a set of events associated with the predilection cohort;

analyzing, by an inference engine, the set of events and description data for the identified member, to generate a predilection score, wherein the inference engine analyzes the set of events and the description data using a rule set, and wherein the predilection score indicates a probability of a future occurrence of the potential action being performed by the identified cohort member; and

responsive to a determination that the predilection score exceeds a threshold, identifying the potential action as an action that is likely to occur.

2. The computer implemented method of claim 1 wherein the set of events comprises at least one of past events, current events, and current circumstances associated with members of the predilection cohort.

3. The computer implemented method of claim 1 further comprising:

receiving a query, by the inference engine, wherein the query comprises an identification of the potential action and the description data for the identified member.

4. The computer implemented method of claim 1 further comprising:

analyzing the set of events and the description data to identify the potential action from a set of potential actions.

5. The computer implemented method of claim 1 wherein the occurrence of the potential action comprises at least one of beginning performance of an action, continuing performance of the action, ceasing to perform the action or refraining from beginning to perform the action.

6. The computer implemented method of claim 1 further comprising:

receiving sensor data in an analog format from a set of multimodal sensors associated with a subject; and

converting the sensor data from the analog format into a digital format and processing the sensor data to form the digital sensor data.

7. The computer implemented method of claim 1 wherein the description data for the identified subject comprises data describing at least one of a previous history of the identified member, a current status of the identified member, an identification data for the identified member.

8. The computer implemented method of claim 1 further comprising:

responsive to a determination that the predilection score fails to exceed a threshold, receiving the digital sensor data associated with the predilection cohort from the set of multimodal sensors and analyzing the digital sensor data for changes in the attributes and changes in the set of events;

responsive to a determination that new digital sensor data associated with the predilection cohort is available, receiving the new digital sensor data, wherein the new digital sensor data comprises updated metadata describing updated attributes associated with the predilection cohort; and

generating an updated predilection score for the predilection cohort based on the description data for the identified member, the set of events, and the updated attributes.

9. The computer implemented method of claim 1 wherein the threshold comprises an upper threshold and a lower threshold, and wherein the predilection score exceeds the threshold if the predilection score exceeds the upper threshold, and wherein the predilection score exceeds the threshold if the predilection score is less than the lower threshold.

10. The computer implemented method of claim 1 wherein receiving the digital sensor data associated with the predilection cohort from the set of multimodal sensors further comprises:

receiving cohort data for a set of multimodal cohorts, wherein the cohort data comprises metadata describing attributes of members of the set of multimodal cohorts; and

generating the predilection score for the predilection cohort based on the set of events, the attributes associated with the set of multimodal cohorts, the attributes associated with the predilection cohort, and the description data, wherein the set of multimodal cohorts comprises at least one of a video cohort, an audio cohort, an olfactory cohort, a biometric cohort, a furtive glance cohort, and a sensor and actuator cohort.

11. The computer implemented method of claim 1 further comprising:

generating the predilection cohort, wherein generating the predilection cohort comprises;

receive manual input from a user; and

analyzing the digital sensor data using the manual input and at least one of cohort criteria, cohort constraints, a set of data models, and sensor patterns, by a cohort generation engine.

12. The computer implemented method of claim 1 wherein the set of multimodal sensors comprises at least one of a set of chemical sensors, a set of audio sensors, a set of cameras, a set of biometric sensors, and a set of sensors and actuators.

13. A computer program product for generating a predilection cohort, the computer program product comprising:

a computer usable medium having computer usable program code embodied therewith, the computer usable program code comprising;

computer usable program code configured to receive digital sensor data associated with a predilection cohort from a set of multimodal sensors, wherein the predilection cohort comprises an identified member of the predilection cohort, and wherein the digital sensor data comprises metadata describing attributes of the identified member;

computer usable program code configured to process and parse the digital sensor data using a set of data models to identify a set of events associated with the predilection cohort;

computer usable program code configured to analyze, by an inference engine, the set of events and description data for the identified member, to generate a predilection score, wherein the inference engine analyzes the set of events and the description data using a rule set, and wherein the predilection score indicates a probability of a future occurrence of the potential action being performed by the identified cohort member; and

computer usable program code configured to identify the potential action as an action that is likely to occur in response to determination that the predilection score exceeds a threshold.

14. The computer program product of claim 13 wherein the set of events comprises at least one of past events, current events, and current circumstances associated with members of the predilection cohort.

15. The computer program product of claim 13 further comprising:

computer usable program code configured to receive a query, by the inference engine, wherein the query comprises an identification of the potential action and the description data for the identified member.

16. The computer program product of claim 13 wherein the description data for the identified subject comprises data describing at least one of a previous history of the identified member, a current status of the identified member, an identification data for the identified member.

17. An apparatus comprising:

a bus system;

a communications system coupled to the bus system;

a memory connected to the bus system, wherein the memory includes computer usable program code; and

a processing unit coupled to the bus system, wherein the processing unit executes the computer usable program code to receive digital sensor data associated with a predilection cohort from a set of multimodal sensors, wherein the predilection cohort comprises an identified member of the predilection cohort, and wherein the digital sensor data comprises metadata describing attributes of the identified member, process and parse the digital sensor data using a set of data models to identify a set of events associated with the predilection cohort, analyze, by an inference engine, the set of events and description data for the identified member, to generate a predilection score, wherein the inference engine analyzes the set of events and the description data using a rule set, and wherein the predilection score indicates a probability of a future occurrence of the potential action being performed by the identified cohort member, and identify the potential action as an action that is likely to occur in response to determination that the predilection score exceeds a threshold.

18. The apparatus of claim 17 wherein the set of events comprises at least one of past events, current events, and current circumstances associated with members of the predilection cohort.

19. The apparatus of claim 17 wherein the description data for the identified subject comprises data describing at least one of a previous history of the identified member, a current status of the identified member, an identification data for the identified member.

20. A data processing system comprising:

a sensor analysis engine, wherein the sensor analysis engine receives digital sensor data associated with a predilection cohort from a set of multimodal sensors, wherein the predilection cohort comprises an identified member of the predilection cohort, and wherein the digital sensor data comprises metadata describing attributes of the identified member;

processes and parses the digital sensor data using a set of data models to identify a set of events associated with the predilection cohort; and

an inference engine, wherein the inference engine analyzes the set of events and description data for the identified member, to generate a predilection score, wherein the inference engine analyzes the set of events and the description data using a rule set, and wherein the predilection score indicates a probability of a future occurrence of the potential action being performed by the identified cohort member, and wherein the inference engine identifies the potential action as an action that is likely to occur in response to determination that the predilection score exceeds a threshold.