Classifying entities by behavior

• US 10,380,486 B2
• Filed: 01/20/2015
• Issued: 08/13/2019
• Est. Priority Date: 01/20/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Krauss, Kirk J.
• Primary Examiner(s): Cao, Phuong Thao
• Application Number:
  US14/600,699
• Publication Number:
  US 20160210317A1
• Time in Patent Office: 1,666 Days
• Field of Search: 707737
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition

Abstract

Methods and apparatus, including computer program products, implementing and using techniques for classifying entities. Data representing an observation of an entity is received. One or more features of the entity are identified. A first classification for the entity is assigned based on the identified features, and a best fit analysis of the identified features to existing classifications. In response to identifying one or more additional features of the entity and determining that the identified additional features meet an update criterion, or in response to determining that the identified features meet an update criterion based on a modified best fit analysis, the assigned first classification for the entity is changed to a second classification for the entity.

Claims

1. A computer-implemented method for classifying physical entities in an entity resolution engine, comprising:

receiving, by a computer, data representing an observation of a physical entity, wherein the data includes a representation of geographic coordinates for the physical entity;

identifying, by the computer, one or more features of the physical entity based on the data representing the observation of the physical entity;

assigning, by the computer, a first classification for the physical entity based on the identified one or more features and a best fit analysis of the identified one or more features to existing classifications, wherein the best fit analysis includes one or more of;

semantic analysis, natural language analysis, image analysis, motion detection analysis, pattern-matching analysis, behavior analysis; and

in response to identifying, by the computer, one or more additional features of the physical entity based on data representing an additional observation of the physical entity and determining, by the computer, that the identified one or more additional features meet an update criterion invalidating assumptions on which the first classification was assigned, or in response to determining, by the computer, that the identified one or more additional features meet an update criterion based on a modified best fit analysis,

defining and adding a second classification to the existing classifications, and

changing, by the computer, the assigned first classification for the physical entity to the second classification for the physical entity.

2. The method of claim 1, wherein the first classification is selected from a set of defined classifications.

3. The method of claim 1, wherein the evidence that the assumption is incorrect includes further observations of one or more physical entities.

4. The method of claim 1, wherein the first classification and the second classification are based on rules and information selected from a group consisting of:

data, location, movement, speed, and behavior patterns pertinent to the observed physical entity.

5. The method of claim 4, wherein the rules are defined programmatically based on observations or resolutions of one or more physical entities.

6. The method of claim 4, wherein the rules are defined by user input.

7. A computer program product for classifying physical entities in an entity resolution engine, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions being executable by a processor to cause the processor to perform a method comprising:

receiving, by the processor, data representing an observation of a physical entity, wherein the data includes a representation of geographic coordinates for the physical entity;

identifying, by the processor, one or more features of the physical entity based on the data representing the observation of the physical entity;

assigning, by the processor, a first classification for the physical entity based on the identified one or more features and a best fit analysis of the identified one or more features to existing classifications, wherein the best fit analysis includes one or more of;

semantic analysis, natural language analysis, image analysis, motion detection analysis, pattern-matching analysis, behavior analysis; and

in response to identifying, by the processor, one or more additional features of the physical entity based on data representing an additional observation of the physical entity and determining, by the processor, that the identified one or more additional features meet an update criterion invalidating assumptions on which the first classification was assigned, or in response to determining, by the processor, that the identified one or more additional features meet an update criterion based on a modified best fit analysis,

defining and adding a second classification to the existing classifications, and

changing, by the processor, the assigned first classification for the physical entity to the second classification for the physical entity.

8. The computer program product of claim 7, wherein the first classification is selected from a set of defined classifications.

9. The computer program product of claim 7, wherein the evidence that the assumption is incorrect includes further observations of one or more physical entities.

10. The computer program product of claim 7, wherein the first classification and the second classification are based on rules and information selected from a group consisting of:

data, location, movement, speed, and behavior patterns pertinent to the observed physical entity.

11. The computer program product of claim 10, wherein the rules are defined programmatically based on observations or resolutions of one or more physical entities.

12. The computer program product of claim 10, wherein the rules are defined by user input.

13. An entity resolution engine, comprising:

a processor; and

a memory storing instructions executable by the processor to perform a method comprising;

receiving, by the processor, data representing an observation of a physical entity, wherein the data includes a representation of geographic coordinates for the physical entity;

identifying, by the processor, one or more features of the physical entity based on the data representing the observation of the physical entity;

assigning, by the processor, a first classification for the physical entity based on the identified one or more features and a best fit analysis of the identified one or more features to existing classifications, wherein the best fit analysis includes one or more of;

semantic analysis, natural language analysis, image analysis, motion detection analysis, pattern-matching analysis, behavior analysis; and

in response to identifying, by the processor, one or more additional features of the physical entity based on data representing an additional observation of the physical entity and determining, by the processor, that the identified one or more additional features meet an update criterion invalidating assumptions on which the first classification was assigned, or in response to determining, by the processor, that the identified one or more additional features meet an update criterion based on a modified best fit analysis,

defining and adding a second classification to the existing classifications, and

changing, by the processor, the assigned first classification for the physical entity to the second classification for the physical entity.